The Cost of NonEurope in The Area of Corruption

The Cost of NonEurope in the area
of Organised Crime
and Corruption
Annex II - Corruption
STUDY
EPRS | European Parliamentary Research Service
European Added Value Unit
PE 579.319 - March 2016
The Cost of Non-Europe in the area of

Organised Crime and Corruption
Annex II- The Cost of Non-Europe in the Area of Corruption
Research paper by RAND Europe
On 7 September 2015, the Coordinators of the Committee on Civil Liberties, Justice and Home
Affairs (LIBE) requested the Directorate-General for Parliamentary Research Services (EPRS) to
prepare a 'Cost of Non Europe Report' on Organised Crime and Corruption to support work on
the own-initiative report on the fight against corruption and follow-up of the CRIM committee
resolution (2015/2110 (INI), Rapporteur Laura Ferrara (EFDD, IT)).
In response to this request, a general assessment1, bringing together the research findings of
three studies commissioned from outside experts, has been drawn up by the European Added
Value Unit of the Directorate for Impact Assessment and European Added Value within DG
EPRS. Its aim is to help improve understanding of the subject matter by providing evidence of
the specific benefits that could be achieved through European action to fight organised crime
and corruption.
The three studies commissioned from outside experts are published as separate documents:
- RAND Europe, research paper on the costs of non-Europe in the area of corruption
(PE 579.319);
- Centre for European Policy Studies (CEPS) & Economisti Associati srl, research paper on the
costs of non-Europe in the area of organised crime (PE 579.318); and
- -Prof. Federico Varese, briefing paper providing an overall assessment of organised crime
and corruption (PE 579.320).
Abstract
Corruption is a phenomenon with significant negative consequences for the EU and its
Member States. This research paper uses a mix of methodologies to quantify the overall costs
of corruption in the EU in economic, social and political terms. The findings, based on new
analysis, suggest that corruption costs the EU between 179bn and 990bn in GDP terms on an
annual basis.
Current anti-corruption measures relevant to Member States and the EU as a whole are
described and their effectiveness in reducing the levels of, and opportunities for, corruption
are assessed. Eight potential areas for EU action are identified that might address the barriers
to the effectiveness of current measures. The costs of non-Europe are calculated in relation to
two of these, as well as in relation to the implementation of recently created EU laws.
Wouter van Ballegooij, Thomas Zandstra, Organised Crime and Corruption: Cost of Non-Europe Report, PE
558.779, European Added Value Unit, March 2016.
1
PE 579.319
Annex II: Corruption
AUTHORS
This study has been written by Marco Hafner, Jirka Taylor, Emma Disley, Sonja
Thebes, Matteo Barberi and Martin Stepanek at RAND Europe and by Professor Mike
Levi at the request of the European Added Value Unit of the Directorate for Impact
Assessment and European Added Value within the Directorate General for
Parliamentary Research Services (DG EPRS) of the General Secretariat of the European
Parliament. This report has been peer-reviewed in accordance with RANDs quality
assurance standards.
RESPONSIBLE ADMINISTRATORS
Wouter van Ballegooij and Thomas Zandstra, European Added Value Unit
To contact the Unit, please email: EPRS-EuropeanAddedValue@ep.europa.eu
LINGUISTIC VERSIONS
Original: EN
DISCLAIMER
The opinions expressed in this document are the sole responsibility of the author and
do not necessarily represent the official position of the European Parliament.
Reproduction and translation for non-commercial purposes are authorised, provided
the source is acknowledged and the publisher is given prior notice and sent a copy.
Manuscript completed in March 2016. Brussels European Union, 2016.
PE 579.319
ISBN 978-92-823-8873-0
doi:10.2861/369191
QA-04-16-192-EN-N
PE 579.319
The Cost of Non-Europe in the area of Organised Crime and Corruption
Table of Contents
Executive summary
1. Quantifying the economic, social and political costs of corruption
in the European Union
2. Gaps and barriers in the existing regulatory framework that
hinder anti-corruption efforts in the European Union
3. Potential policy options within the remit of the LIBE Committee that
might add value and address the challenges identified
4. The costs of non-Europe in corruption
8
9
10
11
13
Acknowledgements
14
Chapter 1 Introduction
1. A study on the potential gains through common action at European
level in the area of corruption
2. Objectives and scope of this paper
3. Research approach and limitations
4. Structure of the paper
15
Chapter 2 Quantifying the costs of corruption

1. The economic, social and political costs of corruption at EU level
2. The costs of corruption in relation to public procurement at EU level
3. Caveats and limitations
25
26
48
60
Chapter 3 The effectiveness of existing anti-corruption measures

1. Description of existing measures
2. Effectiveness assessment
3. Is there potential for action at EU level that might lead to added value to the
challenges identified?
62
63
72
Chapter 4 Costs of Non-Europe: what are the potential benefits of action

at EU-level?
1. Cost of Non-Europe: establishment of a Cooperation and
Verification Mechanism for other Member States
2. Cost of Non-Europe: EPPO
3. Costs of Non-Europe: e-procurement
15
21
22
23
88
103
104
106
110
114
Chapter 5 Report summary and conclusions

1. The costs of corruption in the EU
2. What are the gaps and barriers in the existing regulatory framework
that hinder the effectiveness of measures to combat corruption
in the European Union?
3. Is there a potential for action at EU level and can the gains for
these measures be quantified?
4. Implications and unanswered questions
115
115
References
119
PE 579.319
115
116
117
Table of Tables
Table 1: Summary of EU and international law and monitoring mechanisms
and key enablers and barriers
10
Table 2: Overview of research approach
22
Table 3: Average corruption levels across EU-28 (1995-2014)
33
Table 4: Effect of corruption on GDP per capita and genuine investment
39
Table 5: Group of countries for Scenario 3 (Goodfellas)
41
Table 6: Scenario 1 (The magnificent seven): Average annually reduction in GDP (in US
Dollars)
42
Table 7: Scenario 2 (Catch me if you can) - Average annually
reduction in GDP (in US dollars)
43
Table 8: Scenario 3 (Goodfellas) 44
Table 9: Effect of corruption on inequality, rule of law and organized crime
47
Table 10: Effect of corruption on voter turnout and trust in EU institutions
49
Table 11: 15 NUTS-2 region with highest average CRI index (2009-2014)
54
Table 12: Corruption risk across procurement sub-sectors
55
Table 13: Corruption risk compared across contract with EU funds and no EU funds
56
Table 14: Corruption risk in public procurement and relative contract prices
58
Table 15: Total cost related to corruption risk in EU public procurement
59
Table 16: Summary of EU and international measures to tackle corruption
at the Member State level
63
Table 17: Potential actions at EU level that might lead to added value
to the challenges identified
88
Table 18: Effect of CVM on levels of corruption
105
Table 19 : Potential gains in GDP terms CVM mechanism
106
Table 20: Actions transferred to Member States and subsequent judicial
decisions and convictions before EPPO and after EPPO
108
Table 21: Predicted annual reduction in the costs of corruption after
establishment of EPPO
109
Table 22: Correlation between e-government index (EDI) and public
procurement corruption risk index (CRI)
112
Table 23: Predicted reduction in the corruption levels with the
introduction of e-procurement
112
Table 24: Predicted reduction in the corruption levels with the
introduction of e-procurement
113
Table A-25. Summary of provisions of Article 82 and Article 83 of TFEU
140
Table of Boxes
Box 1: Summary of potential areas for action at EU level in the field of
anti-corruption that might add value and address the challenges of current measures
Box 2: How do our cost of corruption estimates compare to other existing estimates?
Box 3: How does our estimate compare to the PWC and Ecorys (2013) estimate?
Box 4. Overview of key EU-level anti-corruption measures
Box 5: Previous EU legislation relating to public procurement
Box 6: Calls for the protection of whistleblowers from the EU, UN, OECD and CoE
Box 7: Contribution of the Sarbanes Oxley Act to the protection of
whistleblowers in the United States
Box 8: Possible non-legislative action to improve protection for
whistleblowers at Member State level: a whistle-blowing pact
PE 579.319
12
46
60
64
66
77
92
93
Abbreviations
ACR
Anti-Corruption Report
ACWG
Anti-Corruption Working Group
ANTICORRP - FRA project on fundamental rights indicators

BAK
Bundesamt
zur
Korruptionspraevention
und
Korruptionsbekaempgfung (Federal Office of Anti-Corruption)
BEEPS
Business Environment and Enterprise Performance Survey
BPI
Bribery Perception Index
CBA
Centralne Biuro Antykorupcyjne (Central Bureau of AntiCorruption)
CEPEJ
Council of Europe Commission for the Efficiency of Justice
CJEU
Court of Justice for the European Union
CoE
Council of Europe
CPI
Corruption Perception Index
COC
Control of Corruption Index
CPV
Common Procurement Vocabulary (CPV)
CoNE
Cost of Non-Europe
CRIM
Committee
laundering
CVM
Cooperation and Verification Mechanism
EBRD
European Bank for Reconstruction and Development
ECA
European Court of Auditors
ECOSOC
on
organised
crime,
and
money
European Economic and Social Committee
EP
European Parliament
EPPO
European Public Prosecutors Office
EU
European Union
EUJS
European Union Justice Scoreboard
GDP
Gross Domestic Product
GRECO
corruption
Group of States against Corruption
ICRG
International Country Risk Guide
IDEA
International Institute for Democracy and Electoral Assistance
IPP
Institute for Public Policy
JHA
Justice and Home Affairs
LIBE
Civil Liberties, Justice and Home Affairs
MEP
Member of the European Parliament
PE 579.319
MPS
Member of Parliament
NGO
Nongovernmental Organisation
OECD
Organisation for Economic Co-operation and Development
OLAF
European Anti-Fraud Office
OLS
ordinary least squares
PIF Convention
Convention on the protection of the EUs financial interests
OoG
Quality of Government
SCPC
Service Central de Prevention de la Corruption (Central Service

for the prevention of Corruption)
SOCTA
Serious and Organised Crime Threat Assessment
SIENA
Secure Information Exchange Network Application
TI
Transparency International
WEF
World Economic Forum
WGI
World Governance Indicators
WDI
World Bank Development Indicator
UCM
Unobserved Component Model
UN
United Nations
UNCAC
UNDP
UNODC
United Nations Convention Against Corruption

United Nations Development Programme
United Nations Office on Drugs and Crime
VAT
Value-added tax
2SLS
two-stage least squares
PE 579.319
Executive summary
Corruption imposes significant social, political and economic costs on European Member
States and citizens. Corruption defined broadly as 'abuse of power for private gain
can take many forms, including paying bribes or exercising power so as to give
privileged access to public services, goods or contracts. Corruption has been shown to
undermine rule of law, lead to the inefficient delivery of public services and corrode the
institutions and foundations of democracy. Corruption has a measurable impact on
productivity and economic prosperity.
This report looks at the cost of non-Europe in relation to corruption. Cost of non-Europe
reports are intended to study opportunities for gains, or the realisation of a public good,
through common action at the EU level, by attempting to identify areas that might have
large expected benefits as a result of deeper EU integration or coordination. The
objectives of the study are to:
1.
Quantify the economic, social and political costs of corruption in the European
Union.
2.
Investigate gaps and barriers in the existing regulatory framework that hinder the
effectiveness of measures to combat corruption in the EU.
3.
Identify potential for action at EU level that might add value and address the
challenges identified.
This paper focuses on measures for combatting corruption that have been or could be
taken in the field of justice and home affairs, and further focuses on legislative, regulatory
and monitoring measures (rather than soft measures such as awareness raising and
sharing good practices between Member States). It looks at the fight against corruption at
Member State level as well as within EU institutions and in relation to EU funds. Lastly, it
includes a case study focus on the fight against corruption in public procurement.
The data collection methods used to produce this paper consisted of a review of relevant
documents and literature and interviews with 17 stakeholders (including academic
experts in the area of corruption and representatives of EU institutions and agencies).
Additionally, a bespoke data set was compiled from a range of different sources and
using a number of existing corruption indicators, in order to generate new estimates of
the costs of corruption and the costs of non-Europe using econometric modelling.
PE 579.319
1. Quantifying the economic, social and political costs of

corruption in the European Union
Measuring levels of corruption is very challenging. A number of different existing
approaches have been used, including surveys asking about citizens perceptions of the
levels of corruption; surveys asking about actual experiences of corruption; and estimates
based on economic data. Each of these approaches has limitations so that no single
indicator provides a comprehensive picture of the problem. An overview of existing
measures of corruption is provided in Chapter 2.
This study provides a new estimate of the cost of corruption to the EU as a whole,
based on a review of existing literature and econometric modelling. A number of
previous estimates of the costs of corruption have been calculated (see Table A2,
Appendix A). These vary considerably, depending on the data and assumptions on
which they are based.
Our findings, based on new analysis, suggest that corruption costs the EU between
179bn and 990bn in GDP terms on an annual basis.2 These figures are higher than the
estimate of 120bn included in the 2014 EU Anticorruption Report (EC 2014h). The
difference is because the estimate in the EU ACR does not account for the indirect effects
of corruption (it looks at costs in terms of lost tax revenues and foreign investment due to
corruption). Our estimate takes a broader range of effects of corruption into account.
Based on new econometric modelling for this paper we find that corruption in the EU
has significant social costs and political costs. It is associated with more unequal
societies, higher levels of organised crime, weaker rule of law, reduced voter turnout in
national parliamentary elections and lower trust in EU institutions. This finding is in line
with previous research that finds similar relationships between corruption and social and
political cost variables.
Our empirical analysis suggests the cost of corruption risk in EU public procurement
is around 5bn per year (across all sectors). However, the costs of corruption in public
procurement vary considerably between Member States. This estimate is slightly higher
than the estimate provided by a previous, large study.3 This could be because our
estimate includes all sectors of public procurement and all Member States, whereas the
previous estimate included eight Member States and five sectors.
Key limitations to the new estimates produced are outlined throughout Chapter 2.
We calculate three estimates of the costs of corruption in the EU. Each estimate is based on a
different set of assumptions about the extent to which it is feasible for Member States to reduce
corruption in the short, medium and long-term (we refer to these as three scenarios).
3 PWC and Ecorys 2013.
2
PE 579.319
2. Gaps and barriers in the existing regulatory framework

that hinder anti-corruption efforts in the European Union
Anti-corruption policy is primarily a Member State competency the EU does not have
explicit competence in the area of corruption. However, the establishment of the area of
freedom, security and justice is envisaged in the Treaty of Amsterdam (TEU), which
included the fight against corruption as one of its components. Whereas corruption is an
area reserved for Member States, the fight against corruption was one of the foci of the
Stockholm Programme (Council of the EU 2009b) and is explicitly mentioned in a recent
Commission communication, which sets its future agenda in the area of Home Affairs
(COM(2014)154).
A range of measures - including EU and international law and monitoring mechanisms
are in place to reduce, prevent and detect corruption in the EU. These are listed in Table 1
and explained in more detail in Chapter 3, Section I.
This research paper includes an assessment of the effectiveness of these measures. This
assessment was undertaken based on a documentary and evidence review and
interviews with stakeholders. Effectiveness is defined in terms of the ability of the
measures to reduce the level of corruption (or the perception thereof) or to reduce the
opportunity for corruption. Measuring the effectiveness of these measures is very
challenging, given the obstacles to measuring the levels of corruption and any changes
over time. While there are some evaluations of anti-corruption measures, these have not
been conducted in an EU context and their findings do not provide definitive evidence of
what works to address corruption.
Table 1 summarises the key elements found to be effective (enablers) and in need of
improvement (barriers) in relation to each of the key current anti-corruption measures
identified in this paper. These are explained in more detail in Chapter 3, Section II.
Table 1: Summary of EU and international law and monitoring mechanisms and key
enablers and barriers
Type of measure
Enablers
Barriers
EU legislation
e.g. Conventions on the
protection of the EUs
financial interests; Directives
on Public Procurement
- Some legislation has been

transposed by Member
State
- Lack of transposition of
some instruments
- Lack implementation /
enforcement by Member
State
- Some gaps in legislation: no
common protection for
whistleblowers; no common
definition of public official
EU Institutions
e.g. European Anti-Fraud
Office; European Parliament;
European Commission
- Increasingly exercise
oversight in an active
manner
- Provide a good basis for
addressing corruption
- Rely on Member State to

initiate prosecutions
regarding the use of EU
funds
PE 579.319
10
Type of measure
Enablers
Barriers
EU monitoring mechanisms ACR; Cooperation and

Verification Mechanism (for
Bulgaria and Romania CVM); Justice Scoreboard
- ACR raised profile of the

fight against corruption,
produces tailored country
reports using range of
indicators, includes
experience-sharing
programme
- CVM is an important lever to
encourage reform and build
capacity, integrated into
wider reforms, supported by
domestic authorities
- ACR does not cover EU

institutions, does not
generate new data, has no
formal assessment
procedure
- CVM: Mixed evidence of
leading to change, costly to
implement, no strong
sanctions
Council Of Europe
monitoring - the Group of
States against Corruption
(GRECO)
- Employs a systematic
approach and makes good
use of soft enforcement
mechanisms
- GRECO does not cover EU

institutions
3. Potential policy options within the remit of the LIBE

Committee that might add value and address the
challenges identified
Through interviews and a review of previous research and relevant documentation, eight
potential areas for EU action are identified that might address the barriers to the
effectiveness of current measures (summarised in Table 1). These are listed in Box 1,
along with the possible limitations or challenges involved in putting these potential
measures into practice.
The selection of these eight possible actions reflects the remit of this paper, which focuses
on (i) measures that can be taken at the EU level and (ii) hard measures such as changes
to legislation and policy. However, interviewees suggested that, due to limited EU
competencies in the area and the need to work with Member States, some of the most
actionable and feasible policy options may revolve around soft measures (capacity
building, encouragement of political will etc.). Some of these suggested soft measures are
listed in Chapter 3, Section III.
PE 579.319
11
Box 1: Summary of potential areas for action at EU level in the field of anti-corruption
that might add value and address the challenges of current measures
1.
Make use of infringement proceedings against Member States who have not implemented EU law in
relation to the fight against corruption.
Possible challenges/ limitations:

- The number of legislative measures that could serve as the basis for infringement proceedings is
4
limited .
- Might not be as effective as voluntary approaches which are based on encouraging rather than
enforcing compliance.
2.
Support new legislation to harmonise protection for whistleblowing within Member States and/ or
provide protections to whistleblowers within European institutions.

- Lack of a clear legal basis for harmonising Member State legislation (however, the EU has the
competence to act with respect to whistleblowing within EU institutions).
3.
Support new legislation to provide a common definition of a public official.

- Effectiveness would depend on implementation by Member States.
- Proposals for a directive on criminal law protection from fraud and related offences to the EU financial
interests (COM(2012) 363) include a definition of a public official. If this is adopted separate legislation
might not be needed.
4.
Establishment of a European Public Prosecutors Office (EPPO) to address some of the limitations of
OLAF (this has already been proposed and is under discussion).

- There are ongoing negotiations regarding the powers and competencies of the EPPO to operate
somewhat independently of Member States. The outcome of these will determine whether the EPPO
addresses the challenges currently faced by OLAF.
5.
Make improvements to address limitations of the ACR.

- Inclusion of the EU in the ACR would not result in an independent, external review of the EU (because
the ACR is produced by the Commission).
- Increasing the amount of the data included in the ACR or number of outputs per year would require
additional time and resources.
- Replacing the ACR with a broader monitoring framework might not be supported by Member States.
6.
Extend aspects of the Cooperation and Verification Mechanism to other Member States.

- There are questions about whether this would be politically feasible.
- The CVM is designed to be tailored to the needs of the two countries to which it currently applies
(Bulgaria and Romania) and cannot simply be applied to other Member States.
7.
Make improvements to address limitations of the EU Justice Scoreboard.

- A lack of good, comparable data would hinder attempts to expand the Scoreboard to new areas.
8.
Take steps for the EU to accede to GRECO to improve the monitoring of EU institutions.

- Currently some discussion about the ability of the EU, legally, to accede to GRECO.
Although it may be possible to rely on the broader role of the European Commission as the
guardian of the EU Treaties. See further discussion in Chapter 1, Section I (5)
4
PE 579.319
12
4. The costs of non-Europe in corruption

Of the eight potential areas for action at the EU level set out in Box 1, we are able to
provide an estimate of the potential gains from two of them:
-
Establishment of a European Public Prosecutors Office (EPPO) to address some

of the limitations of OLAF.
Extend aspects of the Cooperation and Verification Mechanism to other Member

States.
Unfortunately, it was not possible to undertake quantification of the other potential areas
for action because empirical data about the possible impacts on levels of corruption were
not available.
Additionally, we calculate the potential gains from the adoption of an EU-wide full eprocurement system. Therefore, this research paper provides estimates of the costs of
non-Europe in relation to three policy actions. The aim is to estimate what fraction of the
overall costs of corruption could be recovered with additional action at the EU level.
The estimates of the cost of non-Europe in corruption must be treated cautiously. They
are each based on a set of assumptions that are carefully outlined in Chapter 4 and as
such, should be seen as being indicative of the kinds of gains that might be made. The
estimates look at the potential gains of EU-level action in these three areas, but we do not
look at the costs of implementation of operation of these policies. With these limitations
in mind:
-
We predict that a CVM-like mechanism applied to more Member States, in addition

to Bulgaria and Romania, could reduce the costs of corruption in GDP terms by
around 70bn annually (or around eight per cent of the overall costs).
The establishment of an EPPO could reduce the costs of corruption related to EU

funds by around 0.2bn annually.
The implementation of a full EU-wide e-procurement system could reduce the costs
of corruption risk in public procurement by around 920m each year.
PE 579.319
13
Acknowledgements
The authors wish to thank a number of people for their suggestions and comments on
earlier versions of this document. First of all, we are grateful to Mr Wouter van Ballegooij
and Thomas Zandstra from the Parliamentary Research Service of the European
Parliament for their guidance and suggestions.
In addition, we express our gratitude to all interviewees who kindly agreed to participate
in this study and donated their time and insights.
We are also grateful to a number of experts who reviewed and commented on the
modelling approach in this study. They include Mihaly Fazekas (University of
Cambridge) and Alex Armand (University of Navarra).
We thank Sonja Thebes and Matteo Barberi for excellent research assistance. Finally, we
would like to thank Christian Van Stolk (Rand Europe) and Alex Armand, who peer
reviewed this document as part of RANDs quality assurance process and provided
useful comments and feedback on its earlier versions.
PE 579.319
14
CHAPTER 1 INTRODUCTION
1. A study on the potential gains through common action
at European level in the area of corruption
1.1. Why should we care about corruption?
Corruption is a phenomenon that can and does inflict serious economic, social and
political harms to societies around the world (OECD 2012). The World Bank refers to
corruption as one of the greatest obstacles to economic and social development. It undermines
development by distorting the rule of law and weakening the institutional foundation on which
economic growth depends (World Bank 2009a).
While quantifying the aggregated costs of corruption is not straightforward, a widely
cited estimate by the World Bank calculates that over $1 trillion are paid in bribes in
developing and developed economies each year, which corresponds to about three per
cent of world Gross Domestic Product (GDP) in 2001/2002. Another estimate by the
World Economic Forum (WEF) estimates the cost of corruption to be more than five per
cent of global GDP (US $2.6 trillion) (OECD 2014b).
Besides economic costs, corruption represents a substantial threat as a tool of organised
crime and terrorist groups, often utilised to gain influence and maintain their operations.
The 2013 Serious and Organised Crime Threat Assessment (SOCTA), states that
organised crime groups use corruptive behaviour as a means to infiltrate both public and
private sectors (Europol 2013). The 2014 EU Anticorruption Report (ACR) states that links
between criminal groups, political representatives and businesses are a concern
particularly at the local and regional levels (EC 2014h).
1.2. What are the levels of corruption in European Union Member

States?
Corruption affects all countries in the world, although its effects are most pronounced in
developing countries (UNODC 2004). Measuring levels of corruption is very challenging,
but there are some indicators available. Results from a 2013 special module of the
Eurobarometer survey provide evidence of the extent to which corruption is a perceived
problem in Europe. In that survey (EC 2014f):
Over three quarters of respondents from EU Member States felt that corruption
was widespread in their country.
Approximately a quarter of respondents viewed themselves as being personally

affected (although less than ten per cent reported having experienced or witnessed
corruption personally over the past year) (EC 2014f).
Over half of the respondents believed that corruption in their country has been
constantly increasing in their country. Just five per cent felt it had been decreasing.
PE 579.319
15
The Eurobarometer results show that these perceptions vary considerably across Member
States. For instance, respondents in countries such as Denmark, Finland or Sweden held
relatively positive (i.e. below EU average) views with respect to the pervasiveness of
corruption. By contrast, respondents in new Member States such as Bulgaria, Croatia, the
Czech Republic and Romania expressed much more negative views. 5 This variation is
broadly in line with findings from other cross-national studies on perceptions of
corruption, such as work done by Transparency International (Transparency
International 2012). While there are limitations on using perceptions as proxies for actual
levels of corruption, the link between perceived levels of corruption and decreasing
levels of political trust in institutions means that changing citizens views is an important
concern for policy makers.6
1.3. What does corruption cost the European Union economy?

Table A-2 (in Appendix A) summarises the key existing estimates of the costs of
corruption. The 2014 EU ACR estimates costs to the European Union (EU) economy of
about 120bn a year in lost tax revenues and foreign investment due to corruption. This
EU figure is based on estimates by specialised institutions and bodies, including the
International Chamber of Commerce (ICC), Transparency International (TI), United
Nations Global Compact, the World Economic Forum and Clean Business is Good
Business. A recent study by Mungiu-Pippidi (2013) applied a different approach to
calculating the costs of corruption7 and concluded that the estimate of 120bn in the ACR
underestimates the costs of corruption and that the costs are probably more in the range
of 323bn three times the ACR estimate.
The existing estimates of the costs of corruption each measure slightly different things.
For example, the estimate from the World Bank does not include the embezzlement of
public funds or theft of public assets (which are difficult to measure).8 The World Bank
also acknowledges that its $ 1 trillion estimate of the costs of corruption, does not
account for the significant losses in investment, private sector development, and
economic growth to a country, or to the increases in infant mortality, poverty and
inequality, all resulting from corruption and misgovernance (World Bank n.d.-b). A
similar critique applies to the estimate in the ACR, which includes lost tax revenue and
investments only, not counting further indirect cost components.
99 per cent of respondents in Greece, 97 per cent in Italy and 95 per cent in Lithuania, Spain and
the Czech Republic felt that corruption was widespread in their country.
6 Note that the direction of causality is not a priori clear. It could just be that one is more likely to
believe there is a lot of corruption if one has lost trust in political institutions. We pick up in detail
the issue of reverse causality in Chapter 2. See also a discussion of interpretative limitations
pertaining to data on corruption offered in the section on caveats and limitations later in this
chapter.
7 This study used Denmark, ranked as the least corrupt country in the EU, as a reference point to
estimate the total losses incurred by EU Member States.
8 These types of costs represent financial flows between individuals or from the state to a corrupt
public/elected official. They affect income and wealth distribution, but not necessarily economic
output.
5
PE 579.319
16
The actual costs of corruption are probably higher once indirect effects are taken into
account (OECD 2013b). These indirect effects include changes in the behaviour and
incentives of individuals and firms in light of widespread corruption, which can lead to
lower productivity of labour, physical and human capital. The detrimental effects of
corruption on the efficiency of resource allocation within an economy can operate
through various channels (OECD 2013b):
Corruption weakens market mechanisms. For instance, if investment is subject to

government regulation, then corruption acts as a tax on investments, if businesses
need to pay a bribe in order to get their requests granted. This makes investment
activities more costly for firms, lowering profitability and hence lowering the
overall level of investment.
Corruption has a detrimental effect on competition. This arises in the form of

weaker regulation and antitrust enforcement, as well as deterring new firm entries
into markets and making entrepreneurship less attractive. This matters as
competition is considered as an important driver of efficiency and innovation in
modern economies (i.e. Aghion et al. 2002).
Corruption might affect economic performance by directing the composition and

volume of government expenditures. For instance, the design of the tax system and
its implementation may enable public officials and taxpayers to engage in
corruption activities and therefore lower overall tax revenues.
In summary, when estimating the overall costs of corruption, looking only at specific
areas (i.e. tax revenue) where corruption might have a negative effect, will naturally
underestimate the full costs.
1.4. What is corruption and how do we define it?

For the purposes of this study the definition in the 2014 EU Anticorruption Report is used
as the starting point: 'abuse of power for private gain. This definition allows a relatively
broad scope to examine several forms of corruption (in public and private sectors and by
a range of actors) and the costs of non-Europe in relation to these. 9
However, awareness of other definitions (and disagreements about definition) is a useful
grounding for understanding the nature of the phenomena of corruption to be addressed
by European and national policy makers.10 The following points summarise the key areas
of contention in relation to defining corruption:
This is the definition used by the Global Programme against Corruption, run by the United
Nations (EC 2014h). The other advantage of using this definition is to assist comparability with
other EU studies.
10 As van Stolk and Tesliuc (2010) point out, an important consideration is distinguishing
corruption from fraud and error. In their conceptualisation, fraud is understood as intentional
misleading action to gain a benefit while corruption refers to intentional action by staff or office
holders. The difference between fraud and corruption on one side and error on the other is the
presence or absence of intention. Both corruption and error are likely to be more prevalent in
countries with relatively lower levels of institutional capacity.
9
PE 579.319
17
A focus on public and private actors? A broad working definition was adopted by
the first UN Anti-Corruption Toolkit (2001), which specified that corruption is an
abuse of (public) power for private gain that hampers the public interest (UNODC 2001).
This definition has been criticised for placing too much emphasis on public office
and not enough on the role of private organisations. It has been suggested that a
broader definition is more appropriate in order to encourage private corporations
to share responsibility for tackling corruption (Kaufmann 2005) (but see below for
discussion of the definition of corruption included in a proposed Directive on
Fraud).
A focus on illegal activity? Some definitions attempt to narrow the scope of

corruption to specify that it entails the illegal use of power for personal gain
(underline added) (Zimring and Johnson 2005). This restricts corruption to acts
that that are legally defined as criminal or civil offences. A legal standard (as
opposed to broader standards such as social harm) has the advantage that it is
easier to assess empirically and allows comparability to other related offences
(such as fraud). However, criminal law definitions of corruption focus on specific
types of conduct, which may be too narrow and creates gaps which can be problematic
for non-legal purposes(UNODC 2003), particularly in relation to unethical behaviour
that may not be illegal.
A focus on specified forms of corruption? The Organisation for Economic Cooperation and Development (OECD), the Council of Europe and the UN
Conventions do not provide an overarching definition of corruption and instead
define specific manifestations of the problem. For example, Europol identifies
conflicts of interest, collusion and nepotism as possible forms of corruption and the
Council of Europes Civil Law Convention on Corruption (1999) includes this
specific definition of corruption: Requesting, offering, giving or accepting, directly
or indirectly, a bribe or any other undue advantage or prospect thereof, which
distorts the proper performance of any duty or behaviour required of the recipient
of the bribe, the undue advantage or the prospect thereof(CoE 1999).11
A wider framing of particularism? An approach to understanding and defining

corruption which provides useful framing for this study is the concept of
particularism. This distinguishes individual cases of infringements of norms of
integrity from particularism which is a mode of social organisation characterised
by the regular distribution of public goods on a no universalistic basis that mirrors
the vicious distribution of power within such societies (Mungiu-Pippidi 2006).
Proponents of this definition argue that it is useful for focusing the attention of
policy makers on the root causes of corruption (which can be left untouched by
anti-corruption initiatives that are focused on criminal law interventions), and
draws attention to the stage of development of a particular state or society.
Corruption in developing and post-communist countries is more likely (it is
argued) to constitute particularism.
In respect of this definition it has been commented that a bribe or any other undue advantage or
prospect is broad in scope and may entail a range of activities that might vary across legal, cultural
and geographical contexts and may include bribery of public officials, trading in influence,
embezzlement, misappropriation or other diversion of property by a public official and obstruction
of justice. (See OECD 1997).
11
PE 579.319
18
How is corruption different to fraud? Fraud and corruption are related but
slightly different concepts. Generally, fraud involves a misrepresentation, whereas
corruption is about collusion for gain. The Proposal for a Directive on the fight
against fraud to the Union's financial interests by means of criminal law lists
corruption among fraud-related forms of illegal behaviour and states that
corruption constitutes a particularly serious threat against the Union's financial
interests, which can in many cases also be linked to fraudulent conduct (EC 2012c)
o
Key elements of the proposed Directives definition of corruption are12 the

participation of a public official and the gaining of advantage. The definition
includes passive and active corruption.
The definition of Fraud in the proposed Directive13 does not require the
involvement of a public official (this is in line with distinctions made
elsewhere in available literature, e.g. Van Stolk and Tesliuc 2010), and focuses
on making false, incorrect or incomplete statements in order to
misappropriate funds.
1.5. What are the EU competencies, objectives and actions in the

area of corruption?
As further described in Section II of this chapter, this report looks at the potential gains
through common action at European level in the area of corruption. An important basis
for this is to understand the competence of the EU in relation to corruption.
1.5.1.
Competence stems from establishment of an area of

freedom, security and justice
The competence of the EU in the fight against corruption at the Member State level is
primarily based on the establishment of the area of freedom, security and justice, as
envisaged in the Treaty of Amsterdam (TEU), which included fight against corruption as
Per the directives Article 4: Member States shall take the necessary measures to ensure that the
following conduct, when committed intentionally, is punishable as a criminal offence: (a) the action
of a public official, who, directly or through an intermediary, requests or receives advantages of
any kind whatsoever, for himself or for a third party, or accepts a promise of such an advantage, to
act or refrain from acting in accordance with his duty or in the exercise of his functions in a way
which damages or is likely to damage the Union's financial interests (passive corruption); (b) the
action of whosoever promises or gives, directly or through an intermediary, an advantage of any
kind whatsoever to a public official for himself or for a third party for him to act or refrain from
acting in accordance with his duty or in the exercise of his functions in a way which damages or is
likely to damage the Union's financial interests (active corruption).
13 As stated in the proposals Article 3: any act or omission relating to: (i) the use or presentation of
false, incorrect or incomplete statements or documents, which has as its effect the misappropriation
or wrongful retention or the illegal diminution of funds or assets from the Union budget or budgets
managed by the Union, or on its behalf; (ii) non-disclosure of information in violation of a specific
obligation, with the same effect; (iii) the misapplication of such funds or of a legally obtained
benefit for purposes other than those for which they were originally granted.
12
PE 579.319
19
one of its components.14 Relevant TEU provisions give the EU competence to fight
corruption primarily through the approximation of Member States criminal law and
through measures fostering police and judicial cooperation (Eckes and Konstadinides
2011).
The Treaty of Lisbon (TFEU) did not have a substantial impact on the scope of EU
competency vis--vis tackling corruption. The TFEU provisions pertaining to the area of
freedom, security and justice include Articles 82 and 83,15 which represent the most
important component of the EUs legal basis in the area.
However, as Anagnostou and Psychogiopoulou (2014) point out, the EU does not have
explicit competence in the area of corruption. Similarly, Eckes and Konstadinides (2011)
note that the TFEU does not give the EU any clear competence to initiate common anticorruption standards amongst the Member States anti-corruption policy is dominated
by Member States competencies.
While this is an area reserved for Member States, the fight against corruption was one of
the foci of the Stockholm Programme (Council of the EU 2009b), which guided home
affairs priorities in the European Union between 2010 and 2014. It is explicitly mentioned
in a recent Commission communication, which sets its future agenda in the area of Home
Affairs (EC 2014c). Further, control of corruption is one of the components of Europe
2020, the growth strategy for the European Union covering the current decade (EC 2010),
and is one of the priorities for the Dutch Presidency of the Council from the European
Union from January to July 2016.16
1.5.2.
Routes for addressing corruption stemming from other

parts of the Treaties
In addition to competencies in the area of freedom, security and justice, the EU can draw
on powers in other policy areas, which may be of (indirect) relevance to fight against
corruption at Member State level and at EU level:
For example, the EUs powers to protect its financial interests in accordance with
Article 235 of the TFEU.
Article 114 of the TFEU (related to harmonisation of rules for the establishment
and functioning of the internal market) also provides EU powers relevant to the
fight against corruption, providing a legal basis for EU legislation on public
procurement.
Article 2 TEU. Article 29 TEU (prevention and fight against corruption) was one of the objectives
linked to the establishment of an area of Area of Freedom, Security and Justice (EU 1992).
15 The EU has the potential to establish minimum rules regarding the definition of sanctions and
criminal offences in the areas of serious crime with cross-border dimensions, of which one of them
is corruption (see Article 83 TFEU) (EU 2007).
16 The letter mentions specifically only organized crime. See Minister of Foreign Affairs of the
Netherlands (2015).
14
PE 579.319
20
The role of the Commission as the guardian of the EU Treaties, of which particular
relevance for the fight against corruption is the rule of law principle laid down in
Article 2 of the TEU. Article 7 of the TEU enables the EU to sanction a Member
State found in a serious and persistent breach of values expressed in Article 2.17
However, this procedure is generally considered a nuclear option (Barroso 2012)
and has never been invoked.18
1.5.3.
Mainstreaming the fight against corruption
Mainstreaming refers to the objective of making the fight against corruption an integral
part of EU policies in other related subject areas, so that anti-corruption efforts reflect the
multi-faceted character of the challenge. Mainstreaming recognises that anticorruption
has to be understood in a broader governance context (Mungiu-Pippidi 2013). Examples
of policy areas where mainstreaming is particularly applicable include trade negotiations,
economic growth policy (e.g. European Semester of Economic Governance), public
procurement, general administration, rule of law and others.
2. Objectives and scope of this paper

The European Parliaments Committee on Civil Liberties, Justice and Home Affairs
commenced work on a report on the fight against corruption, representing a follow-up to
the CRIM committee resolution (EP 2015d). Against this background, the European
Parliament has commissioned a series of Cost of non-Europe reports on the topics of
corruption and organised crime. Cost of non-Europe reports are intended to study
opportunities for gains or the realisation of a public good through common action at the
EU level by attempting to identify areas that might have large expected benefits from
deeper EU integration or coordination.
The objective of this study is to assess the current state of play regarding the fight against
corruption in Europe. To that end, the study seeks to:
1.
Quantify the economic, social and political costs of corruption in the European
Union. The study aims to predict in a transparent manner the costs of corruption in
the EU and for all EU Member States, with regard to the overall economic output
lost due to corruption, and not only in specific areas related to foregone tax revenues
or the aggregate value of bribes paid like previous estimates.
2.
Investigate gaps and barriers in the existing regulatory framework that hinder the
effective combat of corruption in the EU. The study seeks to investigate how
effective are different options for combatting corruption at EU level.
The Nice Treaty subsequently amended Article 7 to allow for preventive sanctions as well.
As the Commission pointed out, the Council of Europes Statute includes a similar provision
whereby a Member State found in violation of human rights and rule of law principles may have its
representation suspended and may be expelled. However, like Article 7 of the TEU, this has never
been used (EC 2014a).
17
18
PE 579.319
21
3.
Identify potential for action at EU level that might add value and address the
challenges identified. The study looks at what fraction of the overall cost of
corruption could be reduced by common policy action at the level of the EU.
While corruption covers many different areas relevant to the European policy and
legislative framework, the scope of this research paper is as follows:
Focuses on measures that have been or can be taken in the field of justice and
home affairs (in relation to objectives two and three). This reflects the fact that this
paper is prepared to support the work of the LIBE committee. This means that
measures related to (for example) anti-competition, asset recovery and taxation,
while recognised as important possible levers, are not discussed in this paper.
Similarly, while the importance and desirability of mainstreaming the fight against
corruption in other policy areas is recognised, these efforts remain outside the
scope of this paper. This includes measures related to the regulation of EU funds,
which are also not within the scope of this paper (although it is acknowledged that
financial support and the EUs ability to control and set conditions for its
disbursement represents an enabler and a source of leverage for anti-corruption
measures).
Focuses on possible legislative or regulatory measures (in relation to research

question three), rather than soft measures such as capacity building, awareness
raising and training in order to enhance the fight against corruption.
Includes the fight against corruption at Member State level and within EU
institutions and in relation to EU funds.
Includes a case study focus on the fight against corruption in public

procurement. It was chosen to investigate public procurement in more detail due
to its significance for the EU economy as a whole and its relative high exposure to
corruption risk.
3. Research approach and limitations

The research activities undertaken to produce this research paper are summarised in
Table 2.
Table 2: Overview of research approach
Research Tasks
1
Research methodology
Mapping current state of play and existing

research and information
Document and literature review

19
Interviews with 17 key stakeholders
Interviewees (17 in total) included academic experts and independent researchers in the area of
corruption and representatives of the following organisations: EC DG HOME, EC DG JUST, EC DG
EMPL, EC SG, Europol, Eurojust, EESC, Transparency International, Council of Europe, and
OECD.
19
PE 579.319
22
Research Tasks
2
Quantitative analysis economic, social and

political costs of corruption
Assessing effectiveness of existing

measures, identification of gaps and
barriers, identification of possible areas for
action
Synthesis and analysis
Research methodology
-
Identification and review of relevant

data sets
Build bespoke data set for analysis
Econometric modelling
Thematic analysis of document and
literature review
Internal workshop to review outputs
from tasks 1 and 2
Drawing together findings across all
previous research tasks to produce
research paper
Key limitations to the study are as follows:
Any research into corruption faces challenges in measuring levels of corruption.

By nature an illicit activity, corruption is difficult to detect and measure. Official
statistics and other measures provide only a partial account of the extent and
nature of the phenomena.
There is very limited evidence on the effectiveness of anti-corruption methods.20

Most of the research and evaluation into effectiveness has been conducted in low
and middle-income countries, which might not be transferable to an EU context.
Existing research largely looks at the effectiveness of anti-corruption policy within
a country. This might not provide lessons for the effectiveness of EU action, as the
EU shares competency in the area of fight against corruption with individual
Member States and is heavily reliant on Member States action.
This study was undertaken in a constrained timespan (two months), which

determined the extent of the analysis and the number of stakeholders consulted.
Further limitations of the approach to estimating the costs of corruption and the costs of
non-Europe are discussed in relevant chapters of the paper.
4. Structure of the paper

This research paper is structured as follows:
Chapter 2 sets out existing estimates of the costs of corruption in the EU as a whole
and in relation to public procurement specifically. It sets out new estimates of the
economic, social and political costs of corruption in the EU undertaken by the
research team and new estimates of the costs of corruption in relation to public
procurement.
Although we note there are previous studies. For example: (Hanna et al. 2010; Fjeldstad and
Isaksen 2008; Disch, Vigeland and Sundet 2009).
20
PE 579.319
23
Chapter 3 sets out the current anti-corruption measures relevant to the EU and
Member States and an assessment of the effectiveness of these measures. It sets out
suggestions identified from the review of literature and interviews with
stakeholders for measures that might be taken at the EU level to address gaps
and barriers in existing EU actions.
Chapter 4 sets out quantitative estimates of the Costs of Non-Europe in the area of
corruption.
Chapter 5 provides a report summary and conclusions.
Each chapter of the report commences with a non-technical summary of the research
activities, describes the methodologies and highlights the key findings relevant to the
chapter.
PE 579.319
24
CHAPTER 2 QUANTIFYING THE COSTS OF CORRUPTION
Summary and key findings

Research activities: This chapter addresses research objective 1 by investigating the economic,
social and political costs of corruption at the EU and Member State level and a sub-sector of
the EU economy, public procurement. It starts with a summary of the existing academic
literature and then describes the different corruption indicators and other data sources used in
the course of the analysis. Generally, the study reports estimates for three different corruption
indicators. However, only one indicator (the International Country Risk Guide ICRG - index) is
fully comparable over time, so we focus on this indicator when discussing the implications of
the findings.
The chapter includes a description of the analytical framework (using statistical/econometric
methods) that we used in this paper to assess the costs of corruption within the EU, its
Member States and in public procurement.
Methodologies: Based on a review of existing academic literature (the approach taken to
identify relevant sources is described in Appendix B) and econometric modelling of the
different sorts of harms related to corruption in the EU. The statistical and econometric
modelling techniques are described in detail for the interested reader.
Key findings:
-
We provide three estimates of the costs of corruption in the EU. Each estimate is based on
a different set of assumptions about the size of reductions in the level of corruption that
are feasible for Member States in the short, medium and long-term (we refer to these as
three scenarios). Our findings suggest that corruption costs the EU between 179bn and
990bn in GDP terms on an annual basis (depending on the corruption measure applied).
Unlike some of the existing high-level cost of corruption estimates, these figures include
direct and indirect effects of corruption.
We show that corruption appears to have significant social costs (more unequal societies,
higher levels of organised crime and weaker rule of law) and political costs (lower voter
turnout in national parliamentary elections and lower trust in EU institutions).
Our empirical analysis suggests the cost of corruption risk in EU public procurement is
around 5bn per year overall, including most sub-sectors of public procurement and
contracts of all EU Member States.
PE 579.319
25
1. The economic, social and political costs of corruption at EU

level
1.1. Existing academic research into the costs of corruption
1.1.1. Costs in economic terms
Most empirical research studies that estimate the relationship between corruption and
economic output examine the impact of some aggregate measure of corruption on the level or
growth rate of output (measured usually in GDP terms). The seminal paper by Mauro (1995)
was one of the first empirical studies that looked at the relationship between corruption and
economic development, finding that more corrupt countries experiencing statistically
significantly lower levels of GDP per capita growth and investment rates. Subsequently, a large
number of other studies found very similar results (see, for instance, Mo 2001; Pellegrini and
Gerlagh 2004). Two recent publications (Campos, Dimova and Saleh 2010; Ugur 2014) metaanalyse the large number of empirical estimates that exist in the empirical literature. The
findings of the two meta-studies suggest that a large number of existing estimates indicate a
significant and negative impact of corruption on economic growth and output, but the majority
of estimates suggest a statistically insignificant relationship (around 60 per cent of estimates).
Only a small number of estimates find a positively significant relationship between corruption
and economic output.21 In terms of magnitude of the estimates, the study by Ugur finds a
negative effect of corruption on growth in the magnitude of -0.193, but this estimate is lower (0.079) when low-income countries are excluded, suggesting on average a smaller effect for more
developed countries.
The question that naturally arises is around the direction of causality between corruption and
economic output. Some scholars argue that the causality runs from high corruption levels to
low income (Ehlrich and Lui 1999; Lambsdorff 2007), while some argue that a transition from a
situation with high levels of corruption to one with low levels of corruption is just a by-product
of economic growth (Treisman 2000). One of the mechanisms that have been suggested to
underlie the relationship is that quality institutions are expensive and only rich countries can
afford them. Svensson (2005) gives a summary on this and similar hypotheses and incorporates
the view of a variety of authors with a similar view on the reverse causality hypothesis. In a
recent study, Grundlach and Paldam (2009) demonstrate that in the long-run, causality might
run from low levels of income to high corruption, suggesting that the link between high levels
of corruption and low economic output is part of short to medium-term economic dynamics,
but with higher income over time the quality of institutions (and hence corruption levels)
improve.
In summary, probably both interpretations regarding the direction of causality seem plausible,
nevertheless, the meta-studies by Ugur (2014) and Campos et al. (2010) suggest that causality
runs from corruption to lower economic output and growth. It also depends on the time
horizon over which the impact of corruption is investigated (short-and medium-run vs. the long
run).
21
Often referred to as greasing the wheel corruption.
PE 579.319
26
1.1.2. Costs in social and political terms

Previous research has not only uncovered the costs of corruption in economic terms, but also its
impact on inequality (Gupta, Davoodi and Alonso-Terme 2002; Foellmi and Oechslin 2007),
interpersonal trust (Seligson 2006) or citizens satisfaction with their governments and their life
in general (Tavits 2008). Furthermore, it is well documented in the literature that corruption
negatively affects the good functioning of public institutions and diverts public action from its
intended purpose. For instance, Herzfeld and Weiss (2003) find a strong negative relationship
between corruption and the rule of law. What is more, a number of studies highlight that
corruption and organised crime are not an isolated criminal phenomenon, but that these two
activities reinforce each other (e.g.. Buscaglia 2003; Van Dijk 2007).
In addition to social costs, corruption can also have substantial political costs. A number of
studies find that corrupt and inefficient governments negatively affect electoral participation at
the national and regional level (e.g. Sundstrm and Stockemer 2013), by disrupting the
legislative process and affecting the principles of legality and legal certainty. Corruption can
corrode the institutions and foundations of democracy by producing inefficient delivery of
public services, while grafts and bribes erode the fundamental principles of democracy and
once taken root is likely to spread among institutions (Stockemer, LaMontagne and Scruggs
2013).
It is noted that corruption therefore affects voter turnout. The existing literature provides two
contradictory views on how corruption influences turnout. Firstly, corruption can mobilise
citizens, if citizens prefer clean and accountable governments they turn out in higher numbers if
they do not find transparency and effectiveness in their current government (Inman and
Andrews 2010). Secondly, corruption might repel voter turnout. The mechanism behind that
distrust in the political process alienates citizens from the political process. Disappointment
experienced due to corruption might push potential voters to withdraw from the democratic
process (Kostadinova 2009). Similar to the relationship between corruption and economic
output, the direction of causality is a priori not clear. For instance, Charron and Lapuente (2010)
and Montinola and Jackman (2002) suggest that there might be a reverse causality as democratic
practices might curb corruption (Bck and Hadenius 2008).
Overall, corruption can undermine citizens views and attitudes in political institutions.
Anderson and Tverdova (2003) find that citizens in countries with higher levels of corruption
report more likely negative assessments of the political system and exhibit lower levels of trust
in institutions compared to citizens in countries with lower levels of corruption.
1.2. A brief overview of existing quantitative measures of corruption

Over the last decades, a large number of attempts have been made to measure corruption,
resulting in a number of different corruption indicators and measures that are applied in
research and policy nowadays. The purpose of this section is not to survey all of these
indicators but to discuss the most frequently used and provide a more in-depth discussion
about the indicators used in the remainder of this study.
1.2.1. Perception and non-perceptual indicators of corruption

Due to their illegal nature, corruptive transactions within an economy are clandestine and are
not recorded systematically. Hence, no official statistical aggregate statistics on the extent of the
PE 579.319
27
problem exist and the measures of the extent of corruption can only rely on proxy indicators.
The number of indices focusing on corruption measurement has grown extensively over recent
years and different taxonomies have emerged 22. In essence, these indicators can be classified
into two groups:
1.
Perception-based indicators of corruption, based on surveys or on experts assessments.
2.
Non-perceptual indicators, based on surveys or on evidence-based estimation.
Perception-based indicators are indicators based on the opinions and perceptions of corruption
in a given country; they can draw from a range of surveys among citizens and firms, or from
experts assessments. The most known survey-based composite indicators include the
Corruption Perceptions Index (CPI), the Bribery Perception Index (BPI) and Bribe Payers Index,
published by Transparency International, the Business Environment and Enterprise Performance
Survey (BEEPS) performed by the European Bank for Reconstruction and Development (EBRD)
and the World Bank (WB), and the Corruption Experience Index and the Business International
Index, issued by Business International. Perception measures include as well the indicators of
corruption provided by indices of (global) governance, such as the Corruption Index produced
by the International Country Risk Guide (ICRG), the Bribe Payers Index part of the World
Governance Indicators (WGI) developed by the World Bank, and the Global Competitiveness
index from the World Economic Forum (WEF). The most common non-composite indicators
reporting the original data are the regional barometer surveys (Eurobarometer, Asian Barometer,
Afrobarometer, etc.) that present public opinions towards corruption.
The second group of indicators non-perceptual aim to measure the actual levels of
corruption rather than its perception. Part of these indicators are based on surveys measuring
citizens or firms actual experiences with corruption, such as whether they have been offered,
or whether they have given, a bribe. It is the case of the Global Corruption Barometer (GCB)
developed by Transparency International, where households across the globe report about
experiences with bribes in different forms. Specific Eurobarometer surveys are also designed to
explore the level of corruption experienced by businesses and citizens. Other non-perceptual
indicators seek to provide an evidence-based estimation of the level of corruption using
economic data; such attempts include, for example, the comparison between existing
infrastructure with the total monetary investment in a specific region, and other measures of
missing expenditure. Lastly, there are objective indicators constructed from undisputed facts;
typical examples might include the existence of anti-corruption laws or the funding received by
the anti-corruption agency.
For our quantitative analysis, we ideally apply indicators with values that cover the 28 EU
Member States and a large number of years and not only one or two yearly cross-sections. Such
needs restrict our choices and unfortunately do not allow us to make use of the Eurobarometer
data. Eurobarometer would have represented a useful tool, since these surveys cover in detail
the EU populations perceptions and awareness of the extent of corruption in the Member
States, as wells as direct experiences with corruption and views on the fight against corruption.
The UNDP guide suggests an informal taxonomy that classifies corruption indicators according to scale,
what is being measured, methodology and the role that internal and external stakeholders play; but it also
suggests to distinguish based on the various types of indicators: perception-based indicators and
experience-based indicators, indicators based on a single data source and composite indicators and proxy
indicators (UNDP 2008).
22
PE 579.319
28
However, Eurobarometer surveys on corruption are only available for few years (EC 2014f; EC
2006; EC 2008a; EC 2009; EC 2012b; EC 2014e; EC 2015c). The questionnaires changed
considerably over the years and therefore cannot be used for analyses over time.
Our final choice of indicators has converged on three of the most used indicators in the relevant
economic literature: (1) the corruption perception index (CPI) by Transparency International; (2)
the control of corruption (COC) indicator produced by the Worldwide Governance Indicator
(WGI) project by the World Bank; (3) a corruption index produced by the International Country
Risk Guide (ICRG).
1.2.2. Comparison of corruption indicators

In this sub-section, we provide, i) a description and comparison of the three indices used in the
study, ii) a brief summary of the most common critics, iii) a focus for the EU case and why the
perception indicators are better than their reputation.
Brief description of the three perception corruption indices

The CPI established by Transparency International is a composite indicator used to measure
perceptions of corruption in the public sector. The definition used by TI for corruption is abuse
of entrusted power for private gain, and the CPI focuses on corruption in the public sector, or
corruption that involves public officials, civil servants or politicians. It aggregates perception of
corruption data from international institutions and non-profit organisations. Starting from 2012
just one years data from each data source is included, allowing in this way to compare scores
over time.
The COC index is one of the six indicators of the Worldwide Governance Indicators (WGI)
project developed by the World Bank (WB). The Control of Corruption indicator captures the
perceptions of the extent to which elites and private interests exercise public power for private
gain, including both petty and grand forms of corruption, as well as capture of the state. The
WGI compile and summarise information from over 32 existing data sources that report the
views and experiences of citizens, entrepreneurs, and experts in the public, private and NGO
sectors from around the world, on the quality of various aspects of governance.
The corruption index produced by the ICRG index is an assessment of corruption within the
political system. Financial corruption in the form of demands for special payments and bribes is
relevant for the construction of the ICRG indicator, however major weight is given to actual or
potential corruption in the form of excessive patronage, nepotism, job reservations, favour-forfavours, secret party funding and suspiciously close ties between politics and business. The
ICRG experts collect political information and financial and economic data, converting these
into risk points for each individual risk component on the basis of a consistent pattern of
evaluation.
While the CPI and the COC indices are perception-based measures drawn from a range of
surveys, the ICRG is drawn from an expert assessment. Although they converge on the general
idea of creating better data for better governance by privileging the creation of a composite
measure, the indexes present important conceptual and methodological differences. The first
important difference concerns the original purposes of the different organisations that elaborate
the measures of corruption. The CPI is an ad-hoc measure of corruption aiming to provide data
PE 579.319
29
on extensive perceptions of corruption within countries and raise the public awareness of
corruption. Control of Corruption is instead one of the six indicators necessary to assess a
measure of governance and to establish more effective instruments of government assistance.
Lastly, the ICRG indicator aims to furnish an international clientele (investment firms,
universities, multilateral agencies, transnational firms) with ratings affecting political risk,
economic risk and financial risk for developed, emerging and frontier markets. The way in
which corruption has been defined and conceptualised presents important dissimilarities, as
one can notice in the description of the indices above. Differences exist also between the
strategies of data collection and the aggregation methods. The 2014 edition of the CPI
standardises 12 different data sources to calculate a simple average for each country of all the
available rescaled scores. The COC aggregates 32 different surveys and adopts an Unobserved
Component Model (UCM) in which corruption is approximated as a linear function of the
unobserved corruption in each country and a disturbance term. The ICRG assess political and
economic data to assign specific risks points and convert them in a composite indicator.
Critical assessment of corruption indices

The three indicators have been widely used by political and business leaders to make decisions
and are also widely applied in empirical research about corruption. However, they are subject
to a number of limitations.
The first limit of any indicator derives directly from the lack of a shared definition of what
counts as corruption; the definitions of corruption are either too vague or too biased (Thomas
2010) and specific indicators will inevitably (even if implicitly) reflect particular definitions
(Hawken & Munck 2009). The second source of general critics concerns the risks related to the
technique of gathering and aggregating multiple data: corruption indicators might be overly
complex (Pollitt 2011), and they run the risk of losing conceptual clarity (UNDP 2008). Further
critiques regard the possibility of comparing countries that have different underlying sources of
expert, nongovernmental organisation (NGO) and citizen perceptions (Charron 2016);
aggregate measure of corruptions and yearly standardisation would then not allow an accurate
comparison of countries over time (Andersson and Heywood 2009). Perplexity is also common
about the transparency in collecting information (Rohwer 2009; Christiane 2006), especially
when missing data creates the necessity of devising alternative ways to collect information.
The second form of limitation relates to using perceptions data to capture corruption across
countries. The three proposed indicators measure the perception of corruption and not the
actual level of corruption so the relation between perceived and actual corruption is not
guaranteed. It is now widely acknowledged that perception of corruption is not correlated to
experiences but rather driven largely by outside factors, therefore such indicators are prone to
bias and serve as imperfect proxies for actual levels of corruption (Razafindrakoto & Roubaud
2006; Kurtz & Schrank 2007; Melgar, Rossi & Smith 2010; Heywood & Rose 2014). Perceptions
would then not measure corruption itself but only opinions about its causes and incidence.
Most factors that predict perceived corruption, such as level of economic development, state of
democracy, press freedom and so forth, do not correlate well with available measures of actual
corruption experiences (Treisman 2007). There is evidence that perceptions do not accurately
reflect the actual level of corruption in different regions (Latin America, Mexico, Indonesia,
Russia), while the relation is robust in others (United States) (see Charron 2016 for a review of
the different regional studies). Heywood and Rose (2014) summarise the issue of perception-
PE 579.319
30
based, arguing that perceptions are not necessarily a good reflection of either experience or
reality. Moreover, they showed that there had been virtually no substantive change to two of
the main perception-based measures of corruption (CPI and COC) over a period of more than a
decade, a finding that seriously undermines their utility as analytic tools.
Our study considers the costs of corruption in the EU and therefore our quantitative analysis
will include exclusively the EU Member States. The use of such a relatively homogenous sample
might help to overcome part of the critics when using the described indices of corruption.
Particularly in the EU Member States, political and economic data are available and reliable,
surveys methodology should be consistent over the EU population and issues on obtaining and
collecting information are likely reduced to a minimum. Though, naturally, the linked risk that
indices are a biased measure of perception remain. A series of Eurobarometer studies of the
attitudes of Europeans to corruption (EC 2006; EC 2008a; EC 2009; EC 2012b) showed strong
evidence of the disparity between perception and experiences of corruption. The latest study,
conducted in September 2011, found that a strikingly high proportion of EU citizens (74 per
cent, on average) saw corruption as a major problem in their country, occurring within local
(76 per cent), regional (75 per cent) and national (79 per cent) institutions. Yet, personal
experience of corruption remained strikingly low, with an overall average of just eight per cent
of respondents having been asked to pay any form of a bribe for access to services during the
preceding 12 months.
Despite limitations the corruption indicators are still useful in the EU context
The study by Charron (2016) assesses the relationship between experiences and perceptions of
corruption among citizens and experts exclusively in the European countries. The study is the
first one to offer a systematic analysis of the empirical strength of corruption perception
measures in Europe by using new survey data collected by the author based on 85,000
European respondents in 24 countries. The findings are the following: First, in general, the
results show that there is wide variation among countries in Europe with respect to both
perceptions and actual experience with corruption. Second, in the sample of European countries
analysed, the perceptions of how much ones public sector is plagued by corruption are highly
consistent when comparing samples of those who have, and those who have not, recently
experienced public sector corruption. Third, little evidence is found in support of critics claims
that corruption perceptions are driven by outside noise, at least in the sample European
countries and regions. Given the consistency of corruption perceptions between experts and
citizens, as well as actual reported citizen experiences, the study offers a much less pessimistic
view of existing measures than the prevailing literature. It also offers empirical support for
existing measures in Europe due to the high degree of correspondence found between
perceptions (both expert and citizen) and experience with corruption, at both the national and
regional levels of analysis explored here.
Summary
In summary, while corruption perception indicators may be problematic in the context of
transition and developing countries, the bottom line is that at least in the European context,
they seem to perform much better in measuring actual corruption. Furthermore, even though
they are to some extent different in their methods of preparation, the correlations among the
three indicators, ICRG, COC, and CPI tend to be high. For instance, Svensson (2005) reports a
PE 579.319
31
correlation between the COC Index and the CPI Index of 0.97 and a correlation between the
COC Index or the CPI Index and the corruption scores from ICRG of around 0.75. In our sample
of EU-28 Member States, we find a similar strong correlation between the three indices. The
Pearson correlation coefficient between the COC and the CPI is 0.94, between the COC and the
ICRG 0.85 and between the CPI and the ICRG 0.81.
In the remainder of the study, we report cost estimates related to all of the three corruption
indices. However, because the ICRG index is the only indicator consistently measured, and
therefore comparable, over time, this will be our preferred index. A similar approach has
been taken in the empirical literature (i.e. Aidt 2011).
1.3. Empirical Approach: Data and descriptive evidence on the cost

of corruption
This section describes the data used in the empirical part of the study and provides first
descriptive evidence about a few associations between corruption and different types of costs.
1.3.1. A combination of different data sources

The bedrock of our empirical analysis is a rich data source provided by the Quality of
Government institute of the University of Gothenburg, called the Quality of Government
(QoG) data set (University of Gothenburg, n.d.). The QoG database is a comprehensive onestop collection of a large set of different measures and indicators on quality of government,
public economy, private economy, personal economy, education, health, welfare, judicial,
political system, conflict, civil society and media, drawn from a variety of different validated
sources. In essence, we use the QoG standard data set (time-series version) which has the major
advantage that, in most instances, specific variables are available for a large number of
countries and over time. Overall, a large number of variables used in the analysis are directly
drawn from this comprehensive database. Nevertheless, we supplement the QoG data with
information from other sources, including Eurobarometer (i.e. trust in national and EU
institutions), data on corruption levels (i.e. ICRG corruption indicator), and measures of
organised crime (from the World Economic Forum). Furthermore, to create the growth of
genuine wealth per capita (or genuine investment), as proposed by Aidt (2011), we draw back
on data from the World Bank Development Indicator (WDI) database. By using the QoG data as
the base and supplementing it with information from different sources, including the
corruption index ICRG and data23 on voter turnout across Europe, we create a panel of 28 EU
Member States that covers the years 1995 to 2014.24
1.3.2. Levels of corruption across EU-Member States

Table 3 reports the levels of corruption across EU-28 Member States (average corruption levels
by indicator for the years 1995 to 2014). Across all three corruption indices, we observe a similar
pattern where almost the same set of countries show above EU-28 average levels of corruption
over time. These countries include Bulgaria, Croatia, Cyprus, Czech Republic, Estonia, Greece,
Hungary, Italy, Lithuania, Malta, Poland, Romania, Slovakia and Slovenia.
We use the ICRG index available for the years 1994-2014 available on
http://info.worldbank.org/governance/wgi/index.aspx#doc. (As of 22 February 2016)
24 Note that CPI and COC are available in the QoG database.
23
PE 579.319
32
Overall, we do not find strong empirical evidence of movement over time in the corruption
rankings across Member States. Member States that had high levels of corruption in 1995
continue to do so in 2014.25
Before we deep-dive into the assessment of the costs of corruption at EU and Member State
level, it is useful to provide some preliminary evidence on the relationship between corruption
and its costs in economic, social and political terms.
Figures C-1 to C-3 in Appendix C show the relationship between the three corruption indices
and (log) GDP per capita, the WGI Rule of Law index and voter turnout in national
parliamentary elections (figures are averaged for the period 1995-2004).
Table 3: Average corruption levels across EU-28 (1995-2014)
Country
ICRG Index
CPI Index
COC Index
Austria
0.375
4.024
0.749
Belgium
0.469
4.716
1.167
Bulgaria
0.786
6.892
2.803
Croatia
0.728
6.698
2.632
Cyprus
0.49
5.291
1.437
Czech Republic
0.698
6.454
2.251
Denmark
0.234
2.658
0.154
Estonia
0.6
5.251
1.755
Finland
0.182
2.715
0.219
France
0.5
4.73
1.214
Germany
0.365
4.053
0.741
Greece
0.719
6.765
2.356
Hungary
0.615
6.208
2.118
Italy
0.729
6.595
2.305
Latvia
0.789
6.601
2.508
Lithuania
0.761
6.125
2.401
Luxembourg
0.323
3.597
0.59
Malta
0.583
5.369
1.701
Netherlands
0.302
3.212
0.449
Poland
0.688
6.446
2.179
Portugal
0.495
5.171
1.497
Republic of Ireland
0.708
4.119
0.997
Romania
0.75
7.268
2.852
Slovakia
0.708
6.586
2.349
Slovenia
0.628
5.344
1.662
Spain
0.495
5.185
1.478
Sweden
0.276
2.893
0.335
United Kingdom
0.438
3.645
0.735
EU-28 Average
0.551
5.165
1.558
Notes: Entries in the table show the average values of three corruption indices for each Member State over
the years 1995-2014. The CPI index is on a scale from 1 to 10; the ICRG index on a scale from 0 to 1 and the
COC indicator on a scale from 0 to 5 (originally from -2.5 to 2.5 but we rescaled it to have only positive
values). In bold highlighted are values that are above EU-28 average. For all indicators applies that the higher
the higher the level of corruption.
For instance, the spearman rank correlation coefficient is not lower than 0.88 if we compare the ranking
of countries over time.
25
PE 579.319
33
Interestingly, we observe a negative relationship between all three corruption indices and the
different cost measures under consideration. Figure C-1 suggests a negative relationship
between corruption and (log) GDP per capita, meaning that from a simple statistical point of
view, countries that have higher levels of corruption seem to have lower levels of wealth.
Interestingly, the scatterplot reveals that Luxembourg and Italy are outliers in a sense that they
both are relatively wealthy countries but report relatively high levels of corruption. Figure C-3
highlights that with regard to voter turnout there is relatively more variation (around the linear
line in the graph) across Member States compared to, for instance GDP per capita. This is
probably partially driven by the fact that some countries have compulsory voting laws (e.g.
Belgium and Cyprus) or are small countries (e.g. Malta) where turnout might be naturally
higher. We will take this into account in the multivariate regression analysis that follows by
adjusting for these country-specific factors.
Furthermore, while these scatterplots depicted in figures C-1 to C-3 provide first evidence of the
direction of the relationship between corruption and the outcomes of interest, it is important to
note that the further analysis has to take into account other variables that might affect the link
between corruption and for instance economic output. Not controlling for factors like trade
openness for instance, would result in any predicted parameter estimates of corruption to be
biased.
1.4. Empirical methodology: how to identify the relation between

corruption and its costs?
This section explains in detail the methodological approach taken to estimate the costs of
corruption in economic, political and social terms. In essence, it highlights the econometric
specification and describes the empirical strategy to encounter the prevalent reverse causality
issue when estimating the relationship between corruption and costs.
1.4.1. Econometric specification

To assess the economic, social, and political costs of corruption at EU and Member State level,
we run the following reduced-form regression models:
where
is a set of outcome variables of interest for country i at time t.
represents the level of corruption in any given Member State at a given time (based on three
different cross-national corruption indices).
represents a vector of country characteristics
(i.e. level of human capital, investments or institutional quality), which differs depending on the
outcome variable we investigate. Finally,
denotes a random error term. In this setting, if
corruption has a negative effect on an outcome variable y, then we would expect

In order for the parameter
to be an unbiased estimate of the effect of corruption, econometric
theory presumes that the corruption index
is not correlated with the error term
. However, this is unlikely if we estimate the equation above with empirical data due to the
so-called reverse causality problem which imposes a potential bias on the parameter . For
instance, when we investigate the effect of corruption on economic output, it might be that
PE 579.319
34
corruption affects economic performance, but simultaneously, countries, which perform better,
might also have more resources to fight corruption (and therefore lower corruption levels).
Alternatively, if we look at the relation between corruption and inequality, high levels of
corruption might lead to a more unequal distribution of a countrys wealth or income, but
corruption might at the same time be fostered by higher inequality (e.g. where public servants
income is lower than the average income, public servants might be more prone to asks for
bribes).
1.4.2. Instrumental variable approach

In order to mitigate the reverse causality problem, we apply an instrumental variable
approach, which seeks to find variables that are correlated with corruption but not directly with
(over its relationship with the error term
, the so called exclusion restriction). Previous
research has used a variety of instrumental variables to establish a link with corruption, some of
which have been heavily scrutinised.
For instance, Acemoglu et al. (2000) use colonial settler mortality as an instrument for the
quality of institutions to measure their effect on GDP per capita. The idea behind this is that
European settlers migrated to colonies with lower mortality rates and implemented institutions
close to theirs in the home countries, whereas colonies with higher mortality rates were simply
exploited for the benefit of the home country and left without building strong institutions.
However, Glaeser et al. (2004) criticise this instrument as Europeans also brought their human
capital into the colonies, which is directly related to economic output and therefore settler
mortality is more correlated with human capital than with institutions (noting that institutions
only have second-order effect on economic performance). Other instruments applied include,
for instance, ethnolinguistic fractionalisation under the assumption that there is correlation with
corruption but not with economic output (see Mauro 1995). Furthermore, Hall and Jones (1999)
use distance from the equator as an instrument for social infrastructure, arguing similarly to
Acemoglu et al. (2000) that regions more densely settled by western Europeans were more
strongly influenced by western European values and culture. Both instrumental variables,
ethnolinguistic fractionalisation and distance from the equator, have been heavily criticised
because both are likely directly correlated with economic output and other outputs of interest
such as inequality or the rule of law (and therefore is harming the exclusion restriction for being
a valid instrument). For instance, ethnolinguistic fractionalisation can affect economic output
and voter turnout or inequality directly by creating political instability, or climate (related to
distance from equator) has potentially a direct effect on economic output (see studies by
Easterly and Levine 1997 for more detail; Bloom et al. 1998). Overall, these instrumental
variables are specific to studies that looked at the effect of corruption in the more global context
of corruption (with emphasis on developing countries). Therefore, they may not apply well to a
specific subset of countries, such as the EU-28 Member States, and for that reason, we use a
different set of instruments recently introduced to the literature by Esarey (2015).
In technical terms, the idea behind this instrumental variable approach is to use interactionbased instruments in a setting where for instance two exogenous, non-instrumental variables, w
and v, have a conditional relationship with x, but an unconditional relationship with the
outcome y. In this case, the interaction term wv serves as instrument to identify the effect of x on
y. In essence, the endeavour is to find a variable v whose relation with x depends on w, but
whose relation with y does not depend on w. We follow the approach by Esarey (2015) and use,
based on prior work (Esarey & Chirillo 2013), the fact that the relationship between womens
PE 579.319
35
representation in government and corruption is stronger in the presence of institutions that

foster electoral accountability for corruption. The instruments applied in this case are freedom
of press, presidentialism and personalistic against party-centric electoral rules (in interaction
with womens representation in government).
The overall logic behind this instrumental variable approach is that inside of democratic
countries, certain institutions enhance the mechanism of electoral accountability. Firstly, press
freedom, which works by allowing journalists to expose corruption to the voting public.
Secondly, presidential systems, in contrast to parliamentary systems, normally do not allow noconfidence votes that can hold governments responsible for scandals related to corruptive
activities. Thirdly, personalistic electoral rules, including the degree of political party control
over vote lists, give voters more ability to individually target and punish corrupt officials,
whereas party-centred rules allow responsibility for corruption to be diffused through the
entire party. Individually, these effects might influence any outcome of interest, such as
economic output, but there is reason to assume that conditionally on the proportion of women
in government the influence of these effects diminishes. In short, we use the following three
interaction-based instruments for the level of corruption (in addition to include all four
variables separately in each regression model):
1.
Proportion women in government*press freedom
2.
Proportion women in government*presidentialism
3.
Proportion women in government*personalism.
The proportion of women in government is measured by the share of women in the lower
house of the parliament.26 Freedom House measures freedom of press on a scale of 0 to 100,
whereas higher values correspond less freedom. This variable is also available in the QoG data.
For practical reasons we rescale the variable so that higher values indicate more freedom.
Presidentialism is coded as a binary indicator variable, taking the value 1 if the country has a
presidential system, and 0 otherwise. The variable stems from the Database of Political
Institutions (DPI) by Beck et al. (2001). Personalism is measured by the variable of Johnson and
Wallack (1997), which is on an ordinal scale from 1 to 13, where higher levels indicate that
electoral rules promote personal vote seeking, in contrast to more party-centred vote seeking.
The variable is also included in the QoG data.
Note that we apply ordinary least squares (OLS) and the two-stage least squares (2SLS) method
to estimate the parameters of the various models. 2SLS is a computational method used to
estimate the parameters from the instrumental variable approach. In essence, in the first stage of
the approach, each explanatory variable that is endogenous in the equation of interest is
regressed on all of the exogenous variables in the estimation model, including both exogenous
covariates in the equation of interest and the excluded instruments. The predicted values from
these regressions are then used to replace the endogenous variable (corruption in our analysis)
in the second stage. For more details on the approach, see Angrist & Pischke (2008).
26
As measured by the Inter-Parliamentary Union (IPU) and included in the QoG data set.
PE 579.319
36
1.4.3. A set of different outcome and control variables used in the

analysis
To assess the economic, social and political costs of corruption we look at a variety of different
outcome variables (
), where for each we utilise slightly different control variables (
) in
our econometric specifications outlined in equation (1). Table D-1 D-3 in Appendix D
summarise them in more detail for each category of costs we investigate (economic, social or
political costs).
Firstly, to predict the economic costs we focus on two outcome variables of interest, (log) GDP
per capita and genuine investment. The latter was introduced by Aidt (2010) and is a proxy for
sustainable development, which is concerned with improvements in human welfare. Aidt
(2011) notes that GDP is a flow variable that records the value of goods and services produced
within an economy at market prices in a given year. The problem with GDP as a measure for
output is that it can be increased over time by exploiting an economys capital stocks in terms of
renewable and non-renewable resources or the stock of human capital, without sustainable
savings for future generations. Briefly, genuine investment (or genuine wealth per capita) is
measured as the gross national savings, adjusted by four factors. These factors include (1) a
deduction for the consumption of fixed manufacturing capital; (2) an addition for the value of
human capital investments; (3) deductions for the value of carbon dioxide emissions and the
level of local environmental degradation; and (4) deduction to take into account depletion of
energy, mineral and forest resources. The WDI database includes a measure of adjusted net
savings which is transformed into a growth variable by multiplying the estimate of genuine
investment with a wealth ration parameter and subtracted by the population growth rate (for
more detail see Arrow et al. 2004).
As outlined in Table D-1 in Appendix D, to estimate the relationship between corruption and
(log) GDP per capita we control for similar variables as previous empirical work (i.e. Mauro
1995). We include the level of human capital (operationalised as enrolment in secondary
schooling and average life expectancy at birth), the government share in total output, trade
openness, gross capital formation, value added of the manufacturing and service sector and the
income share of the top 20 per cent as a measure of inequality. In addition, we control for the
level of democracy by using the polity 2 variable (included in QoG data) and the variables
necessary for the instrumental variable approach to work, which include the share of women in
lower parliament, freedom of press, presidentialism, and personalism. A similar set of control
variables is included in the analysis of the relationship between corruption and genuine
investment.
Secondly, to assess the social cost of corruption we focus on a set of different variables (see
Table D-2 in Appendix D). To that end, we investigate whether corruption affects inequality,
measured by the Gini-coefficient and the income share of the richest 20 per cent in a country
(similar to Oechslin and Foellmi 2003). In addition, corruption might undermine the rule of law
of a Member State. We measure rule of law with an indicator provided by the Worldwide
Governance Indicators (WGI) from the World Bank (and included in QoG data). Originally, the
indicator for rule of law lies on the interval between -2.5 and 2.5. For practical reasons we
recode the variable to lie between 0 and 1 with higher values representing a better outcome.
Furthermore, we investigate whether corruption is related to organised crime. We use an
organised crime measure provided by the World Economic Forum (WEF) which asks business
leaders to what extent organised crime (i.e. mafia-oriented racketeering, extortion) imposes
PE 579.319
37
costs on businesses? The variable takes the values on the ordinal scale from 1 (to a great extent)
to 7 (not at all). We note that this is far from a perfect measure to capture the prevalence of
organised crime, but to the best of our knowledge, not many other data sources are available
that measure the extent of organised crime. In addition, the downside of the variable is that it is
only available from 2006 onwards. For all of the four outcome variables measuring social costs,
we employ very similar control variables as in the analysis of the economic costs (as outlined in
Table D-1).
Thirdly, to investigate the political costs of corruption we focus on information of voter turnout
(proportion of eligible adult citizens that cast a ballot in their countrys legislative elections) and
trust in political institutions (see Table D-3 in Appendix D). The Institute for Democracy and
Electoral Assistance (IDEA) provides publicly available data containing a global collection of
voter turnout statistics for national presidential and parliamentary elections since 1945, as well
as European Parliament elections.27 As not every Member State has specific presidential
elections, we only focus on voter turnout at parliamentary and EU Parliament elections. We
include in the model for voter turnout specific political institutional factors including
compulsory voting laws28 and the electoral system type. For the latter, we take data from the
Electoral System Design database,29 which includes indicators on a countrys electoral system.
There is the notion that voter turnout might be higher in a proportional representation (PR)
system (Sundstrm and Stockemer 2013). We therefore capture the influence of the electoral
system on turnout by including two indicator variables, one for PR systems and one for mixed
electoral systems, with the remaining category of majoritarian systems serving as a reference
category. We also include the proportion of people with age above 65 in the general population,
as older people tend to participate more likely in votes and elections.
The Eurobarometer has repeatedly asked European citizens about their trust of different
political institutions, including the European Commission, the European Parliament or the
European Union, as well as the national governments. By using the Eurobarometer micro-data
we create variables that measure the proportion of people that report to trust a particular
institution in each Member State. 30
1.5. The costs of corruption at EU and Member State-level: empirical

results
In this section, we report the findings from the various multivariate regression models we ran
to estimate the costs of corruption in economic, social and political terms.
http://www.idea.int/vt/ (As of 22 February 2016)

We created a binary indicator variable taking the value 1 if a country has compulsary voting laws or 0
otherwise. Member States with compulsary voting laws are Belgium, Cyprus, Greece and Luxembourg.
29 At http://www.idea.int/esd (As of 22 February 2016)
30 http://www.gesis.org/eurobarometer-data-service/data-access/ (As of 22 February 2016)
27
28
PE 579.319
38
1.5.1. Economic costs

Regression results
Table 4 reports the coefficients from the OLS and 2SLS regression models for the effect of
corruption on economic output and growth of genuine wealth per capita. Panel A includes the
estimates for economic output (log GDP per capita). We find a statistically significant negative
effect between corruption and economic output, independent of the estimation method and
corruption indicator used in the model. For instance, using the CPI corruption index, we find
that a one-unit increase of the index reduces GDP per capita by about 4.5 per cent (column 2).
Similarly, a one unit increase in the ICRG index reduces GDP per capita by about 25 per cent
(column 4) and a one-unit increase in the COC index reduces GDP per capita by about nine per
cent (column 6).31 Panel B includes the estimates for genuine investment. We also find a
statistically significant negative effect of genuine investment on corruption (except for the OLS
estimate using the ICRG index, column 3). For instance, the coefficient for the CPI index in
column 2, reveals that one-unit increase in the index reduces genuine investment by about 0.36
percentage points.
Overall, we note that the 2SLS estimates are somewhat larger in absolute value than the OLS
estimates, suggesting that the OLS estimates are biased towards zero, which is consistent with
both reverse causality and measurement error. Furthermore, statistically speaking, our
interaction-based instruments work reasonably well. All estimates pass the weak instrument
test (F-stat > 14) and the p-value of the Hansen J test for over-identification is large as well.
Table 4: Effect of corruption on GDP per capita and genuine investment

(1)
estimation method:
OLS
Panel A: log GDP per capita
CPI index
(2)
(3)
(4)
(5)
(6)
2SLS
OLS
2SLS
OLS
2SLS
-0.1770
-0.2501
(0.078)**
(0.145)*
-0.1203
-0.0949
(0.025)***
(0.043)**
-0.0411
-0.0457
(0.011)***
(0.022)**
ICRG index
COC index
First-stage F-stat:
27.011
Over-id test (p-value):

N:
17.504
0.6264
133
133
27.011
0.6419
140
140
0.6264
140
140
Panel B: genuine investment (growth genuine wealth per capita)

CPI index
ICRG index
-0.1885
-0.3627
(0.068)**
(0.132)**
-0.5537
-2.4987
Bear in mind that the coefficients across the three indices cannot be compared directly in terms of their
magnitude as they are on different scales.
31
PE 579.319
39
estimation method:
(1)
(2)
(3)
(4)
(5)
(6)
OLS
2SLS
OLS
2SLS
OLS
2SLS
(0.512)
(0.955)**
COC index
-0.3789
-0.7646
(0.145)**
(0.273)**
First-stage F-stat:
20.66
30.68
26.97
0.2704
0.1771
0.3145
N:
133
133
140
140
140
140
Notes: panel-robust standard errors in parentheses;*** p<0.001, ** p<0.01, * p<0.010. For the purpose of the
analysis, the sample of 28 Member State has been divided into four cross-sections: 1995-1998; 1999-2002; 20032006; 2007-2010; 2011-2014. All outcome and control variable have been averaged over these sub-periods. OLS =
Ordinary Least Squares; 2SLS = Two-Stage Least Squares IV estimation. All 2SLS estimations have a first-stage Fstatistic > =14. Each model includes time fixed-effects and region fixed-effects (south, east, north; west serves as
reference category).
Predicting the economic costs of corruption in GDP terms

In a next step, we follow a similar approach as Dreher and Hertzfeld (2005) and employ the
estimates included in table 4 to calculate the costs of corruption on the economies of the 28
Member States and the EU as a whole. To that end, we use different scenarios:
1.
Scenario 1 (The magnificent seven): Under this scenario, we benchmark to the average
level of corruption of the seven best performing Member States. Contrary to the approach
taken in Dreher and Hertzfeld (2005) and Mungiu-Pippidi (2013) who benchmark to the
single best performing country, we use a set of different countries. To give an example, we
calculate how much Poland would lose relatively in economic terms by not reaching the
corruption level of the seven best performing Member States. The baseline message behind
the scenario is that some countries do better than others, so at a minimum, how the best
performing countries perform is the least that could be achieved.
By benchmarking to a set of countries, rather than an individual Member State, we
circumvent the fact that, depending on the corruption index, different countries are among
the best performers depending on the index under consideration. By choosing a larger
group as benchmark we ensure that a group of countries is chosen which perform best
across different measures of corruption.
2.
Scenario 2 (Catch me if you can): Under this scenario, we benchmark to the average level
of corruption of the EU-28. Member States above the average corruption level will
converge to the average, whereas Member States below the average do not alter their levels
of corruption. The idea behind this benchmark is that the countries with high levels of
corruption, such as Italy, Bulgaria, Romania or Greece, converge to the average of all
Member States. Hence, this scenario is far less ambitious than assuming that the least
performing countries could perform as good as the seven best performing countries as
under Scenario 1.
3.
Scenario 3 (Goodfellas): Under this scenario, we take a data-driven approach to create

different benchmarks for groups of countries with larger similarities. To that end, we
group Member States into four different groups using a multivariate statistical
PE 579.319
40
approach called cluster analysis.32 The idea behind this is to group countries together
with similar institutional characteristics and levels of corruption. The working
assumption in this scenario is that Member States with higher levels of corruption
converge to the level of corruption of the best performing within their peer group. The
groups of countries resulting from this approach are highlighted in Table 5.
Table 5: Group of countries for Scenario 3 (Goodfellas)

Group
Countries
UK, Germany, Austria, Luxembourg, Ireland
Netherlands, Sweden, Denmark, Finland
Slovenia, Malta, Estonia, France, Belgium, Cyprus, Spain, Portugal
Poland, Hungary, Bulgaria, Greece, Croatia, Latvia, Slovakia, Romania, Italy, Lithuania,
Czech Republic
Within these groups of countries the best performing country serves as benchmark, so for
instance in group 4, if Poland is the best performing country according to a corruption index,
then all the other countries will converge to that level, but not beyond. The advantage of this
scenario is that we compare countries, which have similar institutional settings33, whereas the
downside is that it is also less ambitious in its setup than Scenario 1.
Overall, we see Scenario 1 as our preferred option to benchmark EU Member States in terms of
their foregone economic output in GDP values due to relative high levels of corruption. In the
medium to long-term it probably reflects best what can and should be achieved in terms of
reducing corruption levels within the EU. Nevertheless, in what follows we will also report the
predicted costs of corruption for Scenario 2 and Scenario 3. In essence, in all three scenarios we
calculate the (opportunity) costs of being relatively more corrupt than the sample of benchmark
countries. Note that all three scenarios serve the purpose of a thought experiment, and
especially Scenario 1, might be over-optimistic or unrealistic in their setup, at least in the very
short-term. However, they serve as an illustration of what could be achieved under different
settings.
We use the 2SLS estimates from Table 4 and report the lost output based on all three corruption
indices in Table 6 (Scenario 1), Table 7 (Scenario 2) and Table 8 (Scenario 3).34 Our findings
predict that the costs of corruption in the EU are substantial. For instance, under Scenario 1, the
overall cost of corruption in terms of lost GDP to the EU economy is between $908bn (817bn)
Cluster analysis is an exploratory statistical technique that finds groups of similar observations in
complex data sets. In our application it groups countries by similarities in levels of corruption, the rule of
law and organised crime. The approach consists of two stages: 1) the calculation of a distance matrix
describing differences between variables of interest; 2) creation of clusters based on countries with low
distance to each other. There exist different metrics for distance, we use the Euclidean.
33 Of which some might hinder these countries to unlock their potential for reducing corruption at least in
the short-run.
34 Calculated as the GDP per capita times average population.
32
PE 579.319
41
$1.1 trillion (990bn), depending on the corruption index. 35 This corresponds to about 4.9 per
cent - 6.3 per cent of overall EU-28 GDP. This estimate is in line with the 5 per cent of GDP
estimate made by the WEF.
We further find large Member State variation in terms of the economic costs of corruption, with
the countries with relative high levels of corruption losing most economic output (relative to
their overall GDP). For instance, Bulgaria, Croatia, Romania and Latvia lose almost 15 per cent
of their annual GDP due to corruption. We note that Scenario 1 is probably very optimistic as it
assumes a large reduction the current levels of corruption in many Member States. In our view,
this is still the benchmark to choose as it gives a clear indication of which Member States with
high levels of corruption lose out, at least in the medium or long-term. Nevertheless, we also
provide cost estimates for Scenario 2 (Table 7) and Scenario 3 (Table 8). For instance, under
Scenario 2, were we to assume that all countries above the EU-28 average level of corruption
converge to the average level, we predict an overall cost of above average levels of corruption in
terms of lost GDP of $199bn (179bn) $284bn (256bn) for the EU-28.
For Scenario 3 we grouped Member States with similar institutional settings and corruption
levels. The working assumption of this benchmark is that Member States will at best converge
to the corruption levels of their peers (even though their peers are rather low performing in
terms of their corruption levels) because of institutional settings that may prevent them from
reducing their corruption levels at least in the short-term. Under Scenario 3 we predict an
overall cost of corruption in terms of lost GDP of $242bn (218bn) $314bn (282bn) for the
EU-28.
Table 6: Scenario 1 (The magnificent seven):
Average annually reduction in GDP (in US Dollars)
ICRG index
CPI index
COC Index
Country
Costs (US Dollar)
% GDP
Costs (US Dollar)
% GDP
Costs (US Dollar)
% GDP
Austria
7,456,490,886
2.40%
11,595,343,791
3.73%
8,519,659,975
2.74%
Belgium
19,934,710,965
5.17%
24,134,508,591
6.26%
25,669,982,244
6.66%
Bulgaria
12,707,055,887
14.57%
12,406,069,147
14.22%
19,191,669,331
22.01%
Cyprus
1,654,304,740
5.79%
2,391,053,816
8.37%
2,627,589,371
9.19%
Czech Republic
28,127,023,159
11.95%
29,709,932,588
12.62%
39,624,863,171
16.84%
Germany
63,460,312,004
2.09%
116,472,046,732
3.84%
80,949,879,691
2.67%
Denmark
0.00%
0.00%
0.00%
Spain
74,885,410,532
5.94%
100,564,833,497
7.98%
120,685,426,107
9.58%
Estonia
2,000,127,486
9.05%
1,815,525,498
8.22%
2,690,989,939
12.18%
Finland
0.00%
0.00%
0.00%
France
129,041,552,599
6.10%
133,629,530,786
6.31%
150,448,184,676
7.11%
United Kingdom
83,175,860,582
4.25%
45,849,415,327
2.34%
51,128,831,097
2.61%
Note that the cost estimates are in US Dollars. This allows the comparison with other estimates from the
World Bank and the WEF. To translate these values into EUROS we use an exchange rate of Dollar to Euro
of 0.9.
35
PE 579.319
42
ICRG index
CPI index
COC Index
Country
Costs (US Dollar)
% GDP
Costs (US Dollar)
% GDP
Costs (US Dollar)
% GDP
Greece
34,657,915,647
12.57%
37,956,154,379
13.76%
49,144,572,319
17.82%
Croatia
8,969,660,695
12.83%
9,446,297,322
13.52%
14,262,282,914
20.41%
Hungary
18,061,258,699
9.49%
22,323,155,782
11.72%
29,680,808,592
15.59%
Republic of
Ireland
18,277,528,181
12.26%
6,081,768,710
4.08%
7,559,040,007
5.07%
Italy
257,816,697,848
12.88%
263,112,403,948
13.14%
347,271,680,401
17.34%
Lithuania
6,039,081,396
13.82%
4,990,282,883
11.42%
7,970,506,408
18.24%
Luxembourg
302,697,911
0.86%
763,916,840
2.17%
442,616,165
1.26%
Latvia
4,314,234,558
14.64%
3,877,956,695
13.16%
5,669,094,719
19.24%
Malta
798,249,974
8.56%
806,554,286
8.65%
1,088,284,798
11.67%
Netherlands
1,569,333,137
0.24%
4,910,442,059
0.76%
0.00%
Poland
67,530,248,497
11.64%
73,053,404,311
12.59%
93,690,749,533
16.15%
Portugal
15,187,208,672
5.94%
20,257,458,537
7.93%
24,945,184,033
9.76%
Romania
37,102,972,561
13.49%
42,909,161,118
15.60%
61,799,168,176
22.47%
Slovakia
11,375,652,627
12.26%
12,161,565,453
13.11%
16,468,344,735
17.75%
Slovenia
4,538,632,736
9.88%
3,934,945,726
8.56%
5,197,961,581
11.31%
Sweden
0.00%
0.00%
0.00%
Total EU-28
908,984,221,978
4.91%
985,153,727,819
5.33%
1,166,727,369,983
6.31%
Notes: entries with 0 correspond to the fact that the particular country is below the average benchmark to which costs
are calculated. In Scenario 1, this is the average of the 7 best performing countries. We use the total EU GDP of $18.495
trillion to calculate the share of the cost of corruption among the total EU GDP. Values are in US Dollars.
Table 7: Scenario 2 (Catch me if you can) Average annually reduction in GDP (in US dollars)
ICRG index
Country
CPI index
COC Index
Costs (US Dollar)
% GDP
Costs (US Dollar)
% GDP
Costs (US Dollar)
% GDP
Austria
0.00%
0.00%
0.00%
Belgium
0.00%
0.00%
0.00%
Bulgaria
6,070,081,224
6.96%
5,512,249,461
6.32%
10,178,071,425
11.67%
Cyprus
0.00%
131,967,803
0.46%
0.00%
10,216,549,007
4.34%
11,106,339,000
4.72%
15,300,857,055
6.50%
Germany
0.00%
0.00%
0.00%
Denmark
0.00%
0.00%
0.00%
Spain
0.00%
954,061,381
0.08%
0.00%
Estonia
319,093,223
1.44%
69,436,703
0.31%
407,996,907
1.85%
Finland
0.00%
0.00%
0.00%
Czech Republic
PE 579.319
43
ICRG index
Country
CPI index
COC Index
Costs (US Dollar)
% GDP
Costs (US Dollar)
% GDP
Costs (US Dollar)
% GDP
France
0.00%
0.00%
0.00%
United Kingdom
0.00%
0.00%
0.00%
Greece
13,670,956,818
4.96%
16,157,018,890
5.86%
20,642,428,234
7.49%
Croatia
3,651,134,486
5.22%
3,921,948,874
5.61%
7,039,254,481
10.07%
Hungary
3,571,744,391
1.88%
7,272,910,144
3.82%
10,002,768,153
5.25%
Republic of Ireland
6,931,460,357
4.65%
0.00%
0.00%
105,433,065,735
5.27%
104,831,660,664
5.24%
140,321,246,367
7.01%
2,713,707,104
6.21%
1,536,219,561
3.52%
3,454,354,355
7.91%
0.00%
0.00%
0.00%
Latvia
2,071,947,760
7.03%
1,548,895,392
5.26%
2,623,871,164
8.90%
Malta
88,717,323
0.95%
69,563,352
0.75%
124,676,777
1.34%
0.00%
0.00%
0.00%
23,390,917,839
4.03%
27,205,920,774
4.69%
33,745,639,214
5.82%
Portugal
0.00%
55,798,810
0.02%
0.00%
Romania
16,174,799,666
5.88%
21,171,086,527
7.70%
33,376,860,583
12.14%
Slovakia
4,314,034,395
4.65%
4,826,668,869
5.20%
6,878,043,349
7.41%
Slovenia
1,041,437,673
2.27%
302,412,322
0.66%
448,461,742
0.98%
0
199,659,647,002
0.00%
1.08%
0
206,674,158,526
0.00%
1.12%
0
284,544,529,807
0.00%
1.54%
Italy
Lithuania
Luxembourg
Netherlands
Poland
Sweden
Total EU-28
trillion to calculate the share of the cost of corruption among the total EU GDP.
Table 8: Scenario 3 (Goodfellas) Average annually reduction in GDP due to corruption (in US Dollars)
ICRG index
Country
CPI index
COC Index
Costs (US Dollar)
% GDP
Costs (US Dollar)
% GDP
Costs (US Dollar)
Austria
4,788,568,502
1.54%
10,494,623,405
3.38%
4,618,524,447
1.49%
Belgium
0.00%
2,186,617,155
0.57%
0.00%
Bulgaria
4,434,030,731
5.08%
1,348,020,224
1.55%
5,596,267,137
6.42%
Cyprus
176,123,438
0.62%
763,592,912
2.67%
724,131,236
2.53%
Czech Republic
5,801,519,336
2.46%
0.00%
2,936,445,341
1.25%
Germany
37,408,173,523
1.23%
105,723,563,163
3.48%
42,855,472,000
1.41%
Denmark
3,332,195,897
1.54%
4,961,489,334
2.29%
0.00%
Spain
9,707,378,386
0.77%
28,804,556,860
2.29%
36,755,494,503
2.92%
857,614,190
3.88%
557,631,273
2.52%
1,219,772,419
5.52%
Estonia
PE 579.319
44
% GDP
ICRG index
Country
CPI index
COC Index
Costs (US Dollar)
% GDP
Costs (US Dollar)
% GDP
Costs (US Dollar)
% GDP
Finland
0.00%
4,430,729,792
2.50%
1,075,797,319
0.61%
France
19,565,866,277
0.92%
13,098,037,631
0.62%
9,476,041,719
0.45%
United Kingdom
66,368,279,553
3.39%
38,915,013,046
1.99%
26,552,159,802
1.36%
Greece
8,497,556,876
3.08%
2,989,195,900
1.08%
6,154,179,772
2.23%
Croatia
2,340,088,736
3.35%
584,952,205
0.84%
3,367,634,891
4.82%
Hungary
Republic of
Ireland
0.00%
0.00%
0.00%
16,997,771,411
11.40%
5,553,771,954
3.72%
5,687,731,970
3.81%
Italy
67,869,668,392
3.39%
9,221,776,892
0.46%
35,123,921,605
1.75%
Lithuania
1,893,984,185
4.33%
0.00%
1,158,697,830
2.65%
0.00%
639,030,987
1.81%
0.00%
Latvia
1,519,211,822
5.16%
142,020,013
0.48%
1,075,919,005
3.65%
Malta
316,016,800
3.39%
275,621,013
2.96%
467,311,810
5.01%
Netherlands
22,969,123,022
3.54%
28,020,214,746
4.32%
17,956,870,682
2.77%
Poland
12,510,331,827
2.16%
0.00%
3,274,259,594
0.56%
Portugal
1,968,714,335
0.77%
5,704,045,691
2.23%
7,923,692,450
3.10%
Romania
11,015,890,778
4.01%
8,040,147,526
2.92%
18,929,194,713
6.88%
Slovakia
2,573,306,944
2.77%
396,006,014
0.43%
2,003,088,367
2.16%
Slovenia
2,161,767,524
4.70%
1,318,043,995
2.87%
2,137,265,697
4.65%
9,249,956,236
314,323,138,720
2.77%
1.70%
10,524,644,885
284,693,346,616
3.16%
1.54%
5,646,618,365
242,716,492,674
1.69%
1.31%
Luxembourg
Sweden
Total EU-28
trillion to calculate the share of the cost of corruption among the total EU GDP.
Scenario 2 and Scenario 3 are in a similar ballpark of the overall cost of corruption estimate. The
cost estimate under Scenario 1 is much larger as it assumes a more ambitious reduction in the
corruption levels of the Member States with high levels of corruption.
From a conceptual point of view, Scenario 1 is the preferred option in the view of the authors, as
it is still a more modest benchmark than what other scholars have applied (i.e. best performing
country) and it implies that some countries do better than others, so at a minimum, how the
best performing countries perform is the least that could be achieved, at least in the medium to
long-term, independent of the institutional settings of a particular Member State. We leave it
up to the reader to decide which scenario she prefers. Box 2 provides a discussion on how our
estimates of the cost of corruption compare to other high-level estimates provided by other
institutions.
PE 579.319
45
Box 2: How do our cost of corruption estimates compare to other existing estimates?
Table A2 (Appendix A) gives an overview of existing cost of corruption estimates. How do our
economic cost estimates compare to existing estimates from the EU ACR, the WEF or the
World Bank?
Overall, our number is substantially higher than the 120bn estimate from the European
Commission, especially under Scenario 1 (817 billion 990 billion). This is driven by the fact
that the Commissions estimate includes, to the best of our knowledge, only tax revenues and
EU funds, and does not account for indirect effects of corruption. Our estimate is in GDP terms
and therefore takes, at least to a wider extent, indirect effects into account. Interestingly, our
estimates under Scenario 1 are in line, at least in relative terms, with the WEF estimate, which
postulates that the economic costs of corruption are around five per cent of global GDP.
Our estimate under Scenario 1 is substantially higher than the $1 trillion global value of bribes
paid estimate circulated by the World Bank. However, it is important to note that our
estimates, by measuring the foregone annual GDP, include at least largely the indirect effects
that accrue from high levels of corruption and do not focus solely on specific cost elements
such as bribes paid.
1.5.2. Social costs

Table 9 reports the findings for the effect of corruption on inequality, the rule of law and
organised crime. Similar to previous literature, we find that corruption increases with
inequality. Using the Gini-index as a measure for inequality (higher values equal a more
unequal wealth distribution) our findings suggest that, independent of the corruption index
applied (see Panel A, column 2, 4 and 6 for the 2SLS estimates), we find a positive relationship
between the level of corruption and a larger Gini coefficient. For instance, a one-unit increase in
the CPI index increases the Gini-index by 0.98 points.36
For instance, if Italy reduced the level of corruption (measured by the CPI index, see Table 3) to
the level of Denmark, Italys Gini-index would be lower by 3.86 points. Alternatively, we also
used the share of the 20 per cent richest in a country as different inequality measure and our
findings suggest the same relationship between corruption and inequality as when using the
Gini-index.
Furthermore, Panel B shows the relationship between corruption and the rule of law. Again,
independent of the corruption index used, we find a negative relationship between corruption
and the rule of law. Panel C of Table 2-11 reports similar findings when we look at the
relationship between corruption and organised crime. We therefore find a statistically
significant relationship between corruption, the rule of law and organised crime in the EU.
36
The Gini-coefficient is measured on a scale from 0 to 100.
PE 579.319
46
Table 9: Effect of corruption on inequality, rule of law and organized crime
estimation method:
Panel A: Gini-index
CPI index
(1)
(2)
(3)
(4)
(5)
(6)
OLS
2SLS
OLS
2SLS
OLS
2SLS
0.2731
0.9820
(0.187)
(0.399)**
2.6610
6.4707
(1.527)*
(2.561)**
0.5137
2.4014
(0.479)
(0.886)**
ICRG index
COC index
First-stage F-stat:
17.86
22.667
30.12
0.0882
0.108
0.0821
N:
Panel B: Rule of Law
CPI index
133
133
-0.0660
-0.0573
(0.010)***
(0.024)**
ICRG index
140
140
-0.4759
-0.3526
(0.074)***
(0.158)**
COC index
First-stage F-stat:
17.63

N:
Panel C: Organised Crime
CPI index
133
-0.0408
-0.0430
(0.017)*
(0.025)*
ICRG index
140
-0.1721
-0.1092
(0.020)***
(0.046)**
26.66
0.5216
133
140
27.01
0.5497
140
140
-0.3009
-0.2723
(0.109)**
(0.115)*
COC index
0.3781
140
140
-0.0934
-0.0849
(0.035)*
(0.044)*
First-stage F-stat:
14.24
47.71
18.93
0.8049
0.9225
0.8616
N:
56
56
56
56
56
56
Note: panel-robust standard errors in parentheses;*** p<0.001, ** p<0.01, * p<0.010. For the purpose of the
analysis, the sample of 28 Member State has been divided into four cross-sections: 1995-1998; 1999-2002;
2003-2006; 2007-2010; 2011-2014 when we analyse inequality and rule of law. Organised crime data from
WEF is only available from 2006 onwards, so we split the data into two cross-sections: 2006-2010 and 20102014. All outcome and control variable have been averaged over these sub-periods. OLS = Ordinary Least
Squares; 2SLS = Two-Stage Least Squares IV estimation. All 2SLS estimations have a first-stage F-statistic >
=14. Each model includes time fixed-effects and region fixed-effects (south, east, north; west serves as
PE 579.319
47
1.5.3. Political costs

Table 2-12 reports the effects of corruption on voter turnout and trust in European institutions.
Panel A and B highlight the relationship between corruption and voter turnout at national
parliamentary elections and elections of the European Parliament. Our estimates suggest a
statistically significant negative relationship between corruption and turnout at national
parliamentary elections. For instance, a one-unit increase in the CPI index corresponds to a 0.09
percentage point lower turnout. Interestingly, we do not find such a relationship when we look
at European Parliament elections. The point estimates (columns 2, 4 and 6) are negative, but not
statistically significant. Overall, the average voter turnout is lower at European Parliament
elections (47 per cent) compared to national parliament elections (72 per cent). This could serve
as an explanation for the non-significant effect of corruption on European Parliament elections.
Panels C, D, and E include the estimates for trust in EU institutions. We note that for these
models, our instrumental variable approach was not successful. The F-statistic of the first stage
regressions reveals a weak instrument. This is a technical issue, but from previous estimates, we
would expect that the 2SLS estimates under a non-weak instrument would be larger than the
OLS estimates in magnitude. Therefore, the OLS estimates still provide indication that
corruption is negatively related to trust in EU institutions, just the absolute value of the
parameter is biased towards zero, but statistically significant. Hence, the OLS estimates in
columns 1, 3, and 5 suggest a negative relationship between corruption and trust in EU
institutions, including the European Commission, the European Parliament or the European
Union as a whole. For instance, the OLS estimate in Panel C for the CPI corruption index
suggests that a one-unit increase in the index lowers the share of citizens reporting trust in the
European Commission by 0.08 percentage points.
2. The costs of corruption in relation to public procurement at

EU level
The estimates presented in the previous sections represent the aggregated costs for the whole
EU-economy and its corresponding Member States. However, the costs of corruption can vary
across different sectors in an economy. For instance, public procurement is frequently identified
as a sector vulnerable to corruption, as demonstrated by a section dedicated to this topic in the
2014 EU Anticorruption Report.
As outlined in the first chapter of this study, public procurement has been chosen as a case
study to assess the costs of corruption for a sub-sector of the EU economy, which has a
significant size and has been identified in previous research as one of the sectors with relatively
high levels of corruption risk. Under the term public procurement, we understand the process by
which governments at different levels (national, regional, local) and other public bodies
purchase services, products and public works (EC, n.d.-g).
In this section, we adopt a case study approach and investigate in more detail the costs of
corruption for a specific area, relying on the same two methods as in the previous section, i.e.
document review and econometric modelling. In essence, we revisit the cost estimate by the
PWC and Ecorys (2013) study and based on a different quantitative approach seek to estimate
the costs of corruption (risk) in public procurement for the EU-28 Member States, including all
sectors within the public procurement system.
PE 579.319
48
Table 10: Effect of corruption on voter turnout and trust in EU institutions
estimation method:
(1)
(2)
(3)
(4)
(5)
(6)
OLS
2SLS
OLS
2SLS
OLS
2SLS
-0.2882
-0.6294
(0.123)*
(0.137)***
-0.1239
-0.1842
(0.028)***
(0.035)***
Panel A: Voter Turnout (Parliamentary)

CPI index
-0.0556
-0.0902
(0.014)***
(0.018)***
ICRG index
COC index
First-stage F-stat:
14.69
17.25
15.2
0.9606
0.6062
0.8765
N:
56
56
-0.0266
-0.0242
(0.027)
(0.027)
56
56
0.0619
-0.1796
(0.183)
(0.194)
56
56
-0.0558
-0.0525
(0.054)
(0.054)
Panel B: Voter Turnout (EP)

CPI index
ICRG index
COC index
First-stage F-stat:
14.24

N:
16.28
0.2449
56
18.12
0.2632
56
56
56
-0.7060
-1.6182
(0.249)**
(0.745)*
0.2482
56
56
-0.1955
-0.6223
(0.062)**
(0.276)*
Panel C: Trust European Commission

CPI index
-0.0803
-0.4642
(0.029)**
(0.364)
ICRG index
COC index
First-stage F-stat:
0.312
1.45
1.61
0.0721
0.059
0.0692
N:
56
56
56
56
-0.3009
-0.2723
(0.109)**
(0.115)*
56
56
-0.1519
-0.4731
(0.047)**
(0.232)*
Panel D: Trust European Parliament

CPI index
ICRG index
-0.5243
-1.1255
(0.195)*
(0.622)
COC index
PE 579.319
49
estimation method:
(1)
(2)
(3)
(4)
(5)
(6)
OLS
2SLS
OLS
2SLS
OLS
2SLS
First-stage F-stat:
0.317
1.45
1.611
0.0619
0.031
0.0581
N:
56
56
56
56
56
56
-0.1710
-0.4810
(0.044)***
(0.193)*
Panel E: Trust European Union

CPI index
-0.0738
-0.3438
(0.019)***
(0.264)
ICRG index
-0.6057
-1.4918
(0.171)***
(0.664)*
COC index
First-stage F-stat:
0.317
1.45
1.611
0.0619
0.031
0.0581
N:
56
56
56
56
56
56
Notes: panel-robust standard errors in parentheses;*** p<0.001, ** p<0.01, * p<0.010. For the purpose of
the analysis, the sample of 28 Member State has been divided into two cross-sections: 1996-2005; 20062014. All outcome and control variable have been averaged over these sub-periods. OLS = Ordinary Least
Squares; 2SLS = Two-Stage Least Squares IV estimation. All 2SLS estimations have a first-stage F-statistic >
=14. Each model includes time fixed-effects and region fixed-effects (south, east, north; west serves as
2.1. The nature of corruption in public procurement

Public procurement represents around one third of public spending in developed countries,
therefore corruption in the domain of governmental contracting can have substantial costs
(OECD 2013a). Especially in the EU, public procurement has a strong economic significance,
with around 20 per cent of EU GDP annually spent by government, public sector and utility
service providers for goods, services and public works (PricewaterhouseCoopers and Ecorys
2013).
The World Bank defines corruption in public procurement as an action to steer a contract to the
favoured bidder without detection. This can be handled through different channels, including
(a) avoiding competition through unjustified sole sourcing or direct contracting of awards; (b)
favouring a certain bidder by tailoring specifications or (c) sharing inside information (World
Bank 2009b). This represents restricted and unfair access to public resources (see MungiuPippidi 2006).
2.2. Existing measures of corruption in public procurement and their

corresponding costs
In the infrastructure sector, it is estimated that the financial loss incurred due to corruption
ranges between ten and 30 per cent of publicly funded construction projects (Transparency
International 2005). In that regard, the study by Alexeeva et al. (2011) uses a data set of road
sector contracts funded by the World Bank for 14 countries of Europe and Central Asia covering
the years 2000 to 2014. The authors constructed a list of 11 different red flags, including for
instance cost increases by more than 20 per cent during implementation, time overrun by 30 per
PE 579.319
50
cent, or contract value more than 20 per cent above its engineers estimate, as indicators for
corruption risk among the public procurement contracts analysed. The empirical findings
suggest that one additional red flag is associated with $91,000 to $100,000 cost increase per km
of road rehabilitation and/or road reconstruction. Collier at al. (2015) investigates the cost of
road infrastructure in low and middle-income countries and finds that costs are higher in
countries with higher levels of corruption, measured by the Worldwide Governance Indicators
(WGI). For instance, countries with corruption levels above the median of the WGI measure
have about 15 per cent higher costs. Using a randomised field experiment, the study by Olken
and Barron (2007) investigates road projects in over 600 Indonesian villages. The findings of the
study suggest that increased monitoring of projects (from an average of four per cent to 100 per
cent) reduces (on average) missing expenditures from around 28 per cent to 19 per cent.
From an EU perspective, the study by Fazekas & Tth (2016) shows that relative infrastructure
contract values (ratio between actual contract price and estimated price) increase in every EU
member with a higher risk of corruption (measured as corruption risk based on red flags).
Prices for infrastructure projects with large corruption risk, compared to contracts without
corruption risk are about 16 per cent higher on average across the EU. However, the association
between corruption risk and relative prices of contracts varies across the EU, with countries
such as Poland, Greece and the UK showing a strong relationship between corruption risk and
relative price. For instance, in Poland, if a public procurement contract moves from the lowest
corruption risk to the highest, this would result in about 28 per cent higher infrastructure prices.
A large study by PWC and Ecorys (2013) investigates the costs of corruption in EU public
procurement, with a special emphasis on different sub-sectors (not exclusively infrastructure) of
public procurement involving EU funds. It is noted that EU Structural and Cohesion Funds
contribute heavily to EU public procurement, with a total budget of these funds of around
347bn over the period of 2007 to 2013, corresponding to 0.4 per cent of the EU-27 budget in
2010. Focussing on eight Member States and five sectors where EU funds are spent and using
EU and national procurement data, the study finds significant costs of corruption in EU public
procurement. The direct public loss in corrupt and grey cases (defined as cases that did not
directly lead to a conviction, but where there were indications of corruption risk, measured by
red flags) led to a cost estimate of corruption in the ballpark of 1.4 to 2.2bn in 2010. Using
data for all sub-sectors and Member States from Tenders Electronic Daily (TED) for the years
2009-2014 (EC 2015e), a recent study by Fazekas & Koscis (2015) finds that corruption risks in
public procurement contracts increases the relative price of the contract by around 16 per cent.
We will follow closely the approach by Fazekas & Koscis (2015) to quantify the costs of
corruption in public procurement for all EU Member States and a large sub-set of procurement
sectors.
PE 579.319
51
2.3. Empirical Approach: How to investigate the costs of corruption in

public procurement?
2.3.1. Using a large public procurement database and a measure of
corruption risk
Fazekas & Koscis (2015) develop a comprehensive public procurement database (PP database)
derived from public procurement announcements in 2009-2014 in the EU and Norway.37 The
database includes all public procurement procedures conducted under the EU Public
Procurement Directives. The database was released by DG GROW, which conducted a number
of data quality checks and enhancements. In essence, TED includes a set of variables that
appear in calls for tenders and contract award notices. The TED database contains over 2.8
million comparable contracts for the EU Member States. The PP database includes an indicator
on whether EU funds have been used in public procurement. This information was identified
from each contract award announcement if the use or non-use of EU funds was provided.
Unfortunately, it doesnt breakdown the share of EU funds used for each contract, all that is
known is whether funds have been used in a specific contract or not. In addition, the PP
database includes information on the type and sector of the contracting body, the awarded
contract value, the contracts Common Procurement Vocabulary (CPV) code, as well as the
relative price of the contract (measured as the awarded contract value divided by the estimated
contract value). After some data cleaning, the data includes 575,756 contracts over the time
period 2009 to 2014 for all EU-28 Member States. It is important to note though that for Malta
the total number of contracts is rather low; nevertheless, to capture all 28 Member States we
include Malta as well.
The most important variable included in the PP database is a novel measure of corruption risk
in public procurement. Fazekas & Koscis (2015) construct an objective measure of corruption in
public procurement, which makes use of a range of public procurement red flags, similar to
the flags identified in the PWC and Ecorys study (2013). Fazekas & Koscis (2015) use several
different indications of corruption to create a composite measure called the Corruption Risk
Index (CRI). In essence, the CRI is a weighted measure of the following red flags:
No call for tender published in official journal: for instance, simple way of fixing
tenders is to avoid publication of the call for tenders in official public procurement
journal (i.e. TED) as this makes it harder for competitors to bid.
Non-open procedure type: for instance, invitation tenders are less competitive and use
less open and transparent tender procedures.
Length of advertisement period: for instance when the number of days between
publishing a tender and the submission deadline is short, preparing an adequate bid
might be difficult and hence could serve corrupt purposes.
Weight of non-price evaluation criteria: for instance, different types of evaluation

criteria are more prone to fiddling to different degrees, as subjective and hard to
quantify evaluation criteria could create room for corruptive behaviour.
Length of decision period: for instance, if the time used to decide on the received bids is
very short, this can signal a corruption risk, whereas very short decisions may represent
premeditated assessment.
37
Available at http://digiwhist.eu/resources/data/ (As of 22 February 2016)
PE 579.319
52
The CRI index is a weighted measure of the red flags described above expressed as a scored
between 0 and 1. It is important to stress that the CRI measures corruption risk and not direct
corruptive practices in the procurement of a contract. It is more a warning signal of a potential
corruptive behaviour taking place. At the same time, a low value of the CRI index does not
automatically imply that corruption did not occur.
In what follows, we use the PP database to analyse the extent of corruption risk across the EU,
including NUTS regions and sub-sectors, and to what extent EU funds are associated with
corruption risk. What is more, we seek to calculate the costs of corruption risk in EU public
procurement for all Member States.
2.3.2. Descriptive evidence: Corruption risks across Europe

EU Member States level
The corruption risk in public procurement varies substantially across EU Member States (see
Figure 1). For instance, Luxemburg reports the lowest value for the CRI index and the highest
risk for corruption among public procurement contracts is observed in Croatia. The countries
with above EU-28 average corruption risk in public procurement are Poland, Romania,
Lithuania, Cyprus and Croatia.
Figure 1: Corruption risk across EU Member States
Source: PP database
European NUTS 2 regions

We also observe variation in corruption risks in public procurement across European NUTS-2
regions and not only across EU Member States (see Figure 2). Interestingly we further see
variation across regions within Member States. For instance, within Spain or Italy we observe
some regions with relative high risks in public procurement contracts compared to other
PE 579.319
53
regions. Even within Germany, which across the EU average has relatively low levels of
corruption risks, there is one region that reports a relative high value of the CRI index (DE 71).
The NUTS-2 regions with the highest average risk scores across all procurement contracts are
reported in Table 11. The regions are located in Cyprus, Croatia, Poland, Lithuania, Germany,
and Romania.
Figure 2: Corruption risk across EU NUTS-2 regions
Source: PP database
Table 11: 15 NUTS-2 region with highest average CRI index (2009-2014)
PE 579.319
NUTS-2 Region
CRI Index
CY00
0.590
HR01
0.580
HR03
0.572
PT30
0.566
DE71
0.542
HR02
0.533
LT00
0.532
PL52
0.523
54
NUTS-2 Region
CRI Index
PL43
0.487
RO22
0.482
PL61
0.476
PL33
0.467
RO32
0.463
PL32
0.463
EU-28 average
0.392
Notes: table entries show CRI index values for the NUTS 2 regions with the highest level of corruption risk
(CRI) across public procurement contracts included in TED.
Public Procurement Sub-Sectors (CPV)

The procurement sub-sectors with the highest levels of corruption risk across Europe are
presented in Table 12.
Table 12: Corruption risk across procurement sub-sectors
CPV Code
CRI Index
Collected and purified water
0.801
Recreational, cultural and sporting services
0.711
Public utilities
0.629
Administration, defence and social security services
0.598
Education and training services
0.513
Agricultural, forestry, horticultural, aqua cultural and apicultural services
0.489
Services related to the oil and gas industry
0.488
Postal and telecommunications services
0.475
Supporting and auxiliary transport services; travel agencies services
0.462
EU-28 average
0.392
Source: PP database
Corruption risks and EU funds in public procurement across EU Member States

Fazekas et al. (2013) argue that external funds, such as EU funds, could deteriorate the quality
of government and hence increase the risk for corruption. The authors mention three reasons
for this. Firstly, EU funds are often spent on investment projects where public discretion is
relatively high and it is documented that discretionary spending is more likely related to
corruption than non-discretionary spending. Secondly, EU funds provide a large pool of public
resources for rent extraction of public agents. Thirdly, EU funds weaken the link between
domestic civil society, taxation and policy performance.
Table 13 highlights the fact that across the EU-28 Member States around 15 per cent of contracts
use EU funds and that the corruption risk measured by the CRI index is slightly smaller for
contracts that involve these funds. When we look at the countries where a large share of public
procurement contracts involve EU funds, like Czech Republic, Greece, Lithuania, Luxembourg,
PE 579.319
55
Bulgaria, Portugal or Hungary, the corruption risk in contracts with EU funds is mostly smaller
(except for Lithuania). While there is a corruptive risk for contracts that involve EU funds, the
risk seems to be slightly smaller, generally across the board of all EU Member States.
2.3.3. Methodology to estimate costs of corruption risk in EU public

procurement
Corruptive practices within the public procurement system can drive up prices of projects, not
only in the infrastructure sector, as highlighted in the section above. Usually, projects can be
divided into distinct cost components, including for instance, labour or energy costs, which
would allow the estimation of unit costs of each particular project or contract. However, in the
absence of such detailed information we follow the approach taken by Fazekas & Tth (2016)
and use the measure of prices in form of the relative contract value, which is available in the PP
database.
Table 13: Corruption risk compared across contract with EU funds and no EU funds
Country
% contracts with EU Funds
CRI (no EU Fund)
CRI (EU Fund)
Czech Republic
0.45
0.351
0.323
Greece
0.38
0.329
0.232
Lithuania
0.32
0.484
0.528
Luxembourg
0.29
0.156
0.072
Bulgaria
0.28
0.343
0.283
Portugal
0.28
0.339
0.249
Hungary
0.26
0.272
0.223
Latvia
0.26
0.284
0.231
Malta
0.22
0.404
0.304
Belgium
0.18
0.180
0.145
Estonia
0.17
0.360
0.293
Slovakia
0.14
0.358
0.313
Germany
0.10
0.312
0.316
Poland
0.09
0.429
0.472
Ireland
0.08
0.222
0.243
Spain
0.08
0.252
0.246
Finland
0.07
0.328
0.305
Netherlands
0.07
0.297
0.363
Romania
0.07
0.457
0.335
Slovenia
0.07
0.344
0.376
Cyprus
0.05
0.588
0.427
France
0.05
0.221
0.275
Italy
0.04
0.371
0.361
United Kingdom
0.03
0.304
0.313
Denmark
0.03
0.176
0.360
Austria
0.02
0.240
0.181
PE 579.319
56
Country
% contracts with EU Funds
CRI (no EU Fund)
CRI (EU Fund)
Sweden
0.02
0.148
0.100
Croatia
0.01
0.584
0.500
EU-28 Average
0.15
0.326
0.299
Notes: table entries show average CRI index values for all public procurement contracts across the
different 28 Member States divided by contracts where EU funds have been used and for those without EU
funds. Entries in bold highlight values above average.
The relative contract value (or price) is the ratio of actual contract value divided by originally
estimated contract value and has also been used by Coviello & Mariniello (2014) to estimate the
effect of publicity on the number of bidders, and hence prices, in the public procurement system
of Italy. In essence, this ratio gives a rough estimate of price savings a contract achieved
compared to the initial estimate.
Using the relative price of a contract, we assess the costs of corruption in public procurement at
EU and Member State level by running the following reduced-form regression model using
OLS:
Where rprice represents the relative price of a contract i (ratio of actual contract value divided by
originally estimated contract value) in CPV sector c, in region r, country c and year t. CRI
represents the corruption risk indicator included in the PP database; EUF is an indicator
variable taking the value 1 if the contract included EU funds; X represents a set of control
variables including the type of the contracting body (i.e. national central government, EU
institution), the sector of the contracting body (i.e. general public services, health) and the
awarded contract value. represents sub-sector (CPV) fixed effects;
represents region fixed
effects;
represents country fixed effects and
represents time fixed effects (year). By
including these fixed effects, we ensure that we compare similar public procurement markets
that generally can be divided across different regions and CPV sectors within Member States.
denotes a random error term.
2.3.4. Associations between corruption risk and contract values in EU

public procurement
Table 14 reports the association between the corruption risk indicator and relative prices of
public procurement prices. We find that a one-unit increase of the CRI index raises prices (or
reduces cost savings) on average by about 15 per cent. Fazekas & Koscis (2015) report an
association between the CRI index and relative contract value of similar magnitude. The
empirical findings further suggest that relative contract values are slightly lower for contracts
where EU funds are involved, but the effect is not statistically significantly different from zero.
In column (2) we included an interaction effect of the CRI index and the EU funds indicator
variable to check whether the effect of corruption risk, measured by the CRI index, is different
between contracts with EU funds and contracts without. The parameter estimate for the
interaction effect is very small in magnitude and not statistically significant. We therefore
conclude that the effect of corruption risk on relative contract values is the same across
contracts that have EU funds involved compared to contracts without EU funds.
PE 579.319
57
2.3.5. The costs of corruption in EU public procurement

To calculate the aggregated costs of corruption risk in EU public procurement for each Member
State, we employ the estimate of the relative price increase associated with corruption risk (or
foregone price savings) included in Table 14 together with the average contract value and the
total number of contracts included in the PP database for each Member State. In essence, we
multiply the relative price parameter in Table 14 (column 1) by the average CRI index value in
any given Member State, times the average contract value.
Table 14: Corruption risk in public procurement and relative contract prices
(1)
(2)
estimation method:
OLS
dependent variable:
relative contract value
CRI index
EU Fund
0.1521
0.1521
(0.030)***
(0.030)***
-0.0075
-0.0076
(0.004)
(0.008)
CRI Index * EU Fund
0.0003
(0.017)
Observations
575,756
575,756
R-squared
0.0648
0.0649
Notes: Clustered standard errors in parentheses (country level);*** p<0.01, ** p<0.05, * p<0.10. Regression
models presented in column 1 and 2 include other controls variables such as the type and sector of
contracting body, the contract value and CPV, region, country and year fixed effects. The 95 per cent
confidence interval for the CRI index parameter estimate is [0.091, 0.221].
Multiplied by the total number of contracts we derive the total costs of corruption risk. In
addition, we derive an upper estimate of what proportion of the total costs relates to EU funds
by using the average share of contracts within each Member State that involves these funds. The
costs related to EU funds need to be understood as an upper estimate because the PP database
does not specify what proportion of the full contract was funded by EU funds.
Overall, our calculations suggest that the costs of corruption risk in EU public procurement are
around 5.33bn annually. This figure represents the total costs across all EU-28 Member States
and all sectors included in the PP database (by CPV code). There is substantial variation across
Member States though. The costs of corruption risk is highest in Poland and the United
Kingdom (both above 1bn), followed by Italy (around 0.7 bn) and the Czech Republic
(0.4bn). It is important to stress that Croatia, for instance, has a relative high corruption risk
across its public procurement contracts, but the overall value of the contracts is rather low,
leading to a relatively low total costs of corruption risk, similar to Bulgaria and Cyprus.
Table 15 further includes a column that relates the total costs to the proportion of contract with
EU funds involved. Remember that the PP Database allows us only to identify whether EU
funds have been used, but not their overall share of a contract. Therefore, the estimated cost
figure of 0.6bn related to EU funds needs to be interpreted as the upper bound estimate,
assuming that for each contract the proportion of EU funds was 100 per cent. Box 3 compares
PE 579.319
58
our estimate to the cost of corruption in public procurement estimate by PWC and Ecorys
(2013).
Table 15: Total cost related to corruption risk in EU public procurement
Member
State
Austria
Belgium
Bulgaria
Cyprus
Czech
Republic
Germany
Denmark
Estonia
Spain
Finland
France
Greece
Croatia
Hungary
Ireland
Italy
Lithuania
Luxembourg
Latvia
Malta
Netherlands
Poland
Portugal
Romania
Sweden
Slovenia
Slovakia
United
Kingdom
EU-28 Total
Cost
related to
EUF
(EURO)
year
CRI
% EUFund
Contract
value
(EURO)
0.239
0.019
3,558,479
257
33,158,751
624,013
0.174
0.176
2,796,542
587
43,309,315
7,601,578
0.326
0.284
510,777
546
13,839,039
3,934,711
0.580
0.052
444,383
449
17,590,453
914,129
0.338
0.454
2,843,961
2731
399,601,343
181,299,929
0.313
0.102
950,431
3106
140,316,050
14,373,436
0.182
0.029
3,970,748
233
25,580,679
731,399
0.349
0.167
2,269,548
225
27,078,887
4,513,148
0.251
0.078
2,148,650
3066
251,734,281
19,611,571
0.327
0.075
1,531,914
456
34,727,318
2,600,110
0.224
0.047
1,246,180
1700
72,110,970
3,407,597
0.292
0.377
1,553,876
1117
77,110,065
29,047,130
0.584
0.006
434,902
1916
74,003,746
463,489
0.259
0.265
2,613,523
2362
243,329,958
64,411,883
0.224
0.082
1,564,199
69
3,674,459
301,767
0.371
0.038
2,333,702
5287
696,029,492
26,571,204
0.498
0.315
740,956
210
11,799,625
3,721,276
0.132
0.285
1,946,565
74
2,897,453
826,913
0.270
0.257
554,220
3864
88,031,718
22,595,505
0.382
0.218
778,769
16
706,756
154,037
0.302
0.070
3,343,660
261
40,078,199
2,813,403
0.433
0.094
387,016
55607
1,417,094,143
133,036,685
0.314
0.280
3,116,323
264
39,249,233
10,983,835
0.449
0.068
793,570
6791
368,013,805
25,074,400
0.147
0.017
3,135,868
273
19,118,372
315,523
0.346
0.066
351,424
2132
39,443,543
2,596,065
0.352
0.142
1,414,792
1716
129,980,618
18,496,512
0.305
0.032
11,500,000
1925
1,025,876,827
32,683,112
97,239
5,335,485,098
613,704,360
Contracts
per year
Cost Total
(EURO) year
Notes: the total costs are calculated by multiplying the CRI index point estimate in Table 14, column 1 with the
average CRI risk in the Member State, the average awarded contract value and the total number of contracts per
year. Applying the 95 per cent confidence interval of the point estimate, the total cost estimate lies in the interval of
3.12 - 7.76bn.
PE 579.319
59
Box 3: How does our estimate compare to the PWC and Ecorys (2013) estimate?
As outlined above, to the best of our knowledge, the PWC and Ecorys (2013) study focused on
eight Member States (France, Hungary, Italy, Lithuania, Netherlands, Poland, Romania and Spain)
and five sub-sectors of public procurement, including infrastructure construction, civil
construction (water and urban & utility construction), social employment support and health.
The study estimates costs in EU public procurement to be around 1.4 to 2.2bn. Our estimate is
higher, but includes all Member States and all sub-sectors of procurement included in the PP
database. Table 15 includes and reports all the specific predicted costs to all the Member States
that have been included in the PWC and Ecorys (2013) study. The combined costs of corruption
are around 3.1bn, which is still larger than the 2.2bn, but also includes more sectors. Another
reason for a discrepancy in the cost estimates between our study and the PWC and Ecorys (2013)
study might be that the latter uses confirmed cases of corruption and grey cases (those with a
large number of red flags) whereas our study employs a novel composite corruption risk index
which is based on a number of such flags. Overall, the real cost estimate therefore may lie
somewhere in between the 1.4 to 2.2bn and 5.3bn. Nevertheless, it is important to stress
that the corruption risk indicator may also pick up non-corruptive cases.

There are significant challenges surrounding the measurement of the extent of the corruption
problem and of the effectiveness of anti-corruption efforts, particularly when it comes to crossnational comparisons.
Criminal justice statistics are one logical source of information, however:
The availability of information on corruption, including data on non-criminal cases, is

uneven across the Member States (EC 2014h). Some cases that contain an element of
corruption may be prosecuted as fraud/falsification of accounts or as organised crime/
drugs/people trafficking offences (or the corruption may lead them not to be prosecuted
for any offences unless the corruption charges are removed).
Where there are data available, they often lack coherence across Member States, not least
because of the definition of corruption and the range of offences that may be conceived as
corruption-related (including by bodies collecting statistics).
The interpretation of criminal justice statistics is not straightforward. For instance, a high
number of corruption cases appearing in courts may suggest a high prevalence of
corruption in a given country. However, this may also be an indicator of more positive
features, such as an environment in which law enforcement and the judicial system have
the necessary tools to combat corruption and thus bring cases to court, or an environment
in which there is a will to report corruption on the part of the general population.
Furthermore, an alternative source of information is surveys that ask about respondents

perceptions of and personal experiences with corruption. Key examples include the
PE 579.319
60
Eurobarometer surveys (as referenced above in Chapter 2 Section I(2), surveys by the World
Bank (n.d.-a), World Economic Forum (n.d.), Transparency International (n.d.) and others.
However, these sources also pose interpretation problems albeit their weaknesses are different
to those related to criminal justice statistics on corruption:
Surveys may be subject to social desirability bias (particularly in so far as questions on

personal experience with corruption are concerned).
Their cross-national application may be affected by differences in national contexts.38
Survey results may be influenced by some of the mechanisms described above. For
instance, if a countrys authorities are successful in bringing more corruption cases to
court this may result in greater visibility and media coverage of corruption, fuelling
perceptions of high prevalence of the phenomenon.
Information on the fight against corruption (although not the extent or cost of corruption) can
be accessed through the standing review mechanisms described above, most notably the work
of GRECO (CoE, n.d.-e), the OECD Working Group on Corruption and UNCAC. These review
mechanisms focus on measures to fight corruption and their implementation, but have limited
ability to provide information on the results of these efforts. In this sense, they are institutional
and activity assessments rather than effectiveness assessments.
RAND has previously studied the cross-national use of survey tools in our work on Understanding
Intolerance in Western Europe (Rubin et al., 2014) For theoretical work regarding possible respondent biases,
see, for instance, (Paulhus 1984; Podsakoff et al. 2003; Tourangeau & Smith 1996).
38
PE 579.319
61
CHAPTER 3 THE EFFECTIVENESS OF EXISTING ANTICORRUPTION MEASURES
Summary and key findings

Research activities: This chapter responds to research objective 2: Are there gaps and barriers
in existing EU actions, including legislation and its implementation which hinder the effective
combat against corruption in the EU? and to part of objective 3 How effective are different
options for combatting corruption at EU level? Is there a potential for action at EU level that
might lead to added value to responses to the challenges identified? The chapter starts by
mapping existing EU actions that apply at Member State level. It then provides an assessment
of the effectiveness of existing anti-corruption measures. Lastly, it identifies areas for potential
EU action.
Methodologies: The chapter is based on a review of existing literature (the approach taken to
identify relevant sources is described in Appendix B) and interviews with experts and
professionals in the field of the fight against corruption.
Key findings:
Key anti-corruption measures applicable to Member States and EU institutions include:
The EU Legislative framework; EU institutions with relevant powers (such as the European
Anti-Fraud Office OLAF); EU Monitoring mechanisms (EU ACR, Cooperation and
Verification Mechanism, EU Justice Scoreboard); Legislation and guidance from
international institutions (Council of Europe, Organisation for Economic Co-operation and
Development and United Nations).
Barriers to the effectiveness of EU or international law stem from a lack of transposition
by Member States or lack of implementation or enforcement at the Member State level.
These issues may be a reflection of differences in factors such as political will or
administrative capacity.
Some gaps in legislation are identified, such as the lack of an EU-wide system of
whistleblower protection and the absence of a harmonised definition of a public official.
Monitoring mechanisms are generally considered to have contributed to the effectiveness
of the fight against corruption at Member State level, although areas for improvement are
identified, including the exclusion of EU institutions from these mechanisms..
Eight potential areas for EU action are identified, to address barriers to the effectiveness
of current measures:
Make use of infringement proceedings against Member States who have not
implemented EU law in relation to the fight against corruption.
Support new legislation to harmonise protection for whistleblowing within Member
State and/ or provide protection to whistleblowers within European institutions
Support new legislation to define a public official.
PE 579.319
62
Make improvements to the ACR report, including extending coverage to the EU

institutions.
Extend aspects of the Cooperation and Verification Mechanism to other Member
States.
Make improvements to the EU Justice Scoreboard.
Take steps for the EU to accede to GRECO.
Establish a European Public Prosecutors Office.
1. Description of existing measures

In order to address research objective 2 (Are there gaps and barriers in existing EU actions,
including legislation and its implementation which hinder the effective combat against
corruption in the EU?), it is necessary to first map measures that have already been passed to
prevent, detect, monitor, etc., corruption. This section provides an overview of measures
intended for implementation at the national level stemming from: the EU; GRECO; OECD; UN
and other sources. The measures discussed in this chapter are summarised in Table 16 .
This chapter is not intended to provide a comprehensive account of all actors and measures in
the fight against corruption. Those discussed have been identified from the literature review
and stakeholder interviews. The discussion of the EUs current action is presented in two
separate sub-categories: 1) legislation and policy, and 2) monitoring and reporting.
Table 16: Summary of EU and international measures to tackle corruption at the Member
State level
Type of measure
EU legislation
EU monitoring
mechanisms
Council of Europe
GRECO
OECD
UN
Scope of
application
All Member
State
ACR: All
Member State
CVM: BG, RO
EUJS: All
Member State
All Member
State
23 Member
State
Is it legally
binding?
Yes
All Member
State
For legally binding measures, summary of

compliance / transposition across Member State
Some gaps in transposition
More serious gaps in implementation
No
Yes
Yes
Yes
A small number of Member State have not

ratified key documents.
Implementing legislation has been passed in
all Member State (five non-OECD EU
Member State not covered)
Member State participation complete assessment of implementation is hampered
by delays in review
ACR - EU Anticorruption Report; CVM - Cooperation and Verification Mechanism; EUJS EU Justice Scoreboard
PE 579.319
63
1.1. EU Legislative framework

In its role of coordinating and supporting the fight against corruption, the EU has developed its
own instruments (as well as relying on work undertaken by other relevant bodies). Box 4 lists
the key EU measures and the date they were adopted.
Box 4. Overview of key EU-level anti-corruption measures
1995
Convention on the protection of EUs financial interests, accompanied by two

protocols on corruption and on money laundering offenses
1997
Convention on the fight against corruption involving EU or EU Member State officials

(entered into force in 2005)
2003
Council framework decision on combatting corruption in the private sector
2008
Accession of the EU to the UN Convention on Corruption
2008
Establishment of the European contact-point network against corruption
2011
EC communication on the state of play in the fight against corruption in the EU
2012
EC proposal for a directive on fight against fraud to the Union's financial interests by
means of criminal law
2014
Revised public procurement directives (on the award of concession contracts, on

public procurement, on procurement by entities operating in the water, energy,
transport and postal services sectors, and on electronic invoicing in public
procurement)
2014
Publication of the EU Anticorruption Report
2015
Anti-money laundering directive
EU legislation related to corruption dates from the late 1990s. The Convention on the protection
of the EUs financial interests (also referred to as the PIF Convention) 39 was adopted in 1995
along with its two protocols on corruption and on money laundering offenses (Council of the
EU, 1995; Council of the EU, 1996; Council of the EU, 1997). A 1997 Convention against
corruption involving EU or EU Member State officials aimed to strengthen judicial cooperation
among Member States on corruption involving European and national officials (Official Journal
C 195, 25/06/1997). This Convention came into force in 2005.
In 2003, the EU adopted legislation addressing corruption in the private sector (2003/568JHA),
which formed the basis for an EU legal framework in the area. In 2008, the EU acceded to the
United Nations Convention on Corruption (UNCAC) (2008/801/EC). That same year, the
Council decided (2008/852/JHA) to establish a contact-point network against corruption, which
would serve as a forum for exchanging information on measures and experience in the fight
against corruption.
39
From the French protection des intrts financiers.
PE 579.319
64
In 2011, the European Commission, in its communication on the state of play in the fight against
corruption in the EU (EC 2011d), announced the creation of an expert group and a biennial EU
ACR, intending to provide a hitherto non-existent assessment of EU Member State efforts to
combat corruption.40 The first (and so far only) EU ACR was published in February 2014.
In other recent developments, the Union undertook a revision of public procurement directives
(previous measures on public procurement are summarised in Box 5) (EU 2014b; EU 2014c) and
adopted a new directive on confiscation of criminal assets (EU 2014d) and on money laundering
(EU 2015a).
In 2012 the Commission put forth a proposal for a directive against fraud to the Unions
financial interests by means of criminal law (EC 2012c). These are still under discussion in the
final stages of agreement.
In 2014 the EU also adopted Directive 2014/55/EU which aims to facilitate the use of einvoicing in Europe and is intended to result in the introduction of new rules to extend the use
of e-procurement (EU 2014e). It also adopted Directive 2014/23/EU on the award of concession
contracts (EU 2014a),41 which addresses a gap in the regulatory framework by offering a precise
definition and coverage for high value works and services concessions.42 Specific changes in the
revised directives of particular relevance to fight against corruption include:
-
Stronger measures intended to prevent, identify and remedy conflicts of interest,

favouritism and corruption.
A clearer notion of conflict of interests.
Provisions allowing anyone attempting to influence a public purchaser or making false

statements to be excluded from public procurement procedures.
Compulsory exclusion from public procurement procedures in cases where there are
abnormally low tenders due to non-compliance with EU law in the field of social and
labour law, environmental law, international social and environmental law.
A monitoring and assessment mechanism had existed prior to the creation of the EU Anticorruption
Report in the case of Bulgaria and Romania, which had been subject to the Cooperation and Verification
Mechanism (EC, n.d.-i). However, as noted in EC (2011d), there had been no mechanism in place
monitoring the existence, and assessing the effectiveness, of anti-corruption policies at EU and Member
State level in a coherent crosscutting manner
41 Concessions are partnerships between the public sector and mostly private companies, where the latter
exclusively operate, maintain and carry out the development of infrastructure (ports, water distribution,
parking garages, toll roads) or provide services of general economic interest (energy, water distribution
and waste disposal, for example).
42 While public contracts were already exhaustively regulated in the 2004 Directives, they only partially
covered works concessions and completely excluded service concessions. According to an impact
assessment carried out by the Commission in advance to the proposal for this directive, the rules and
practices of Member States concerning the award of concessions were very different to other types of
contracts. Moreover, many EU Member States were found not to have rules on concessions at all (EC
2011b).
40
PE 579.319
65
Box 5: Previous EU legislation relating to public procurement
The current European legal framework for public procurement in the EU builds on the
Directives 2004/17/EC (EU 2004a) and 2004/18/EC (EU 2004b) adopted on 31 March 2004,
which were fully transposed by all the Member States by 2010 (EC 2011c). A comprehensive
evaluation by the European Commission conducted in 2011 found that the 2004 Public
Procurement Directives had contributed to greater openness and transparency, leading to
increased competition and savings (EC 2011c). However, it also noted differences in the
implementation and application of the directives across Member States and raised some
concerns about the efficiency of the EU public procurement regime. In other assessments on
the state of play in public procurement, the Commissions 2012 Annual Public Procurement
Implementation Review found e-procurement was used in only five per cent to ten per cent of
procurement procedures carried out across the EU (EC 2011a).
1.2. EU institutions with relevant powers

1.2.1. European Anti-Fraud Office
The key institution to detect and combat corruption at the EU level, damaging the EUs
financial interests, is OLAF (European Anti-Fraud Office). Established in 1999, OLAF has the
authority to investigate two types of cases internally within EU institutions and bodies (EP
and Council of the EU, 1999), and externally in individual Member States (Council of the EU,
1995). Furthermore, OLAF lends support to other EU institutions, notably the European
Commission, in the process of policy development. In 2013, the OLAF Regulation was revised
to strengthen the institutions governance, to bolster its rights in both internal and external
investigations, and to improve its information exchange with other EU institutions and Member
States (EC 2013c).
In addition to OLAF, the fight against corruption at the EU level is supported by a range of
other institutions.
1.2.2. European Parliament

The European Parliament plays a role through its oversight powers on the EU budget. The EP
can submit formal questions and request clarifications from other EU institutions, particularly
the European Commission,43 and exercises financial scrutiny of EU institutions.
Following the Lisbon Treaty, the legislative procedure was amended to incorporate the
principle of qualified majority voting and with an increased role for the European Parliament in
accordance with Articles 223 to 234 and 314 of the TFEU. Within the EP, responsibility for
policies concerning the areas of freedom, security and justice rests with the LIBE committee,
whose portfolio includes measures relating to police and judicial cooperation in criminal
However, a TI evaluation (2014b) notes that there are no sanctions for non-compliance on the part of
those approached by the EP with a request for information.
43
PE 579.319
66
matters, including terrorism, and substantive and procedural measures relating to the
development of a more coherent Union approach to criminal law (EP 2015i). However, given
the relevance of other policy areas to the fight against corruption (see the discussion of
mainstreaming, above) additional standing committees of the Parliament have a role to play in
shaping the EUs response to corruption, such as the Budget Committee (BUDG), Budgetary
Control Committee (CONT), Legal Affairs Committee (JURI) and Constitutional Affairs
Committee (AFCO).
The European Parliament has contributed to the decision making process, for instance in the
form of a resolution with recommendations on action and initiatives to be taken in the areas of
organised crime, corruption and money laundering (EP 2013).44
1.2.3. European Court of Auditors, European Ombudsman and Court

of Justice of the European Union
All EU institutions are regularly audited by the European Court of Auditors, investigated for
maladministration where appropriate and following complaints from the public by the
European Ombudsman, and are subject to judicial oversight by the Court of Justice of the
European Union (CJEU).
1.2.4. Europol and Eurojust

The EUs law enforcement agencies, Europol and Eurojust also play a role in the detection and
investigation of corruption cases at EU level and involving the EUs financial interests (Council
of the EU 2009a). To that end, they have in place cooperation agreements with OLAF and make
use of the SIENA (Secure Information Exchange Network Application tool) (Europol, n.d.), run
by Europol and intended to facilitate rapid and effective information exchange between
European agencies, Member States and relevant third parties.45 However, several interviewees
pointed out that corruption cases represent only a small fraction of the workload of Europol
and Eurojust.46
1.2.5. European Commission

Moving beyond detection and investigation, the Commission has an important enforcement
authority at its disposal. Since December 2014,47 the EC is able to initiate infringement
In March 2011, the European Parliament set up a special committee on organised crime, corruption and
money laundering (CRIM), which was operational for 18 months. The standing committee covering the
relevant areas under Justice and Home Affairs policies is the Committee on Civil Liberties, Justice and
Home Affairs (LIBE).
45 We note a revision of Europols mandate has recently been agreed (EP 2015c).Reflecting on this
development, De Capitani (2015) noted that this revision is not based on any legally-binding framework
for police cooperation. Instead, it utilises soft policy tools, including programmatic documents such as the
European Internal Security Strategy.
46 One interviewee noted, however, that in some instances this may be a result of prosecutorial decisions.
Given the difficulties with proving corruption, prosecutors may decide to press other types of charges,
even though the crime in question involved corrupt behaviour.
47 In accordance with Article 10 of Protocol No 36 on Transitional Provisions of the Treaty of Lisbon.
44
PE 579.319
67
procedures against Member States for failing to transpose EU measures adopted under the third
pillar, i.e. in the area of police and judicial cooperation in criminal matters.48
1.3. EU monitoring mechanisms

The EU manages or supports a set of reporting and monitoring mechanisms, designed to collect
information and indicators on the progress in the fight against corruption and to identify areas
in need of attention.
1.3.1. EU Anticorruption Report

The EUs flagship monitoring project is its ACR. Issued first in 2014 and intended to be
published biennially thereafter, the report draws on existing data sets to monitor trends and to
identify weaknesses. It consists of 28 country-specific reports and a summary report that offers
thematic analyses. The first edition in 2014 was supported by a dedicated expert group on
corruption convened by the European Commission (EC, n.d.-h) and a network of national
correspondents bringing together NGOs and independent advisers, with the aim to minimise
the need to rely exclusively on data from national authorities (Pop 2011).49 The preparation of
the second iteration of the ACR, scheduled to be published in 2016, is in progress. One of the
most important data sets informing the ACR and other publications is a set of surveys
administered as part of the Eurobarometer series, referred to in Chapter 1 and 2 of this report. 50
1.3.2. Cooperation and Verification Mechanism

Another key monitoring mechanism is the Cooperation and Verification Mechanism (CVM).
The mechanism was established following the accession to the EU by Bulgaria and Romania in
recognition of the persistence of certain weaknesses in the area of judicial reform, fight against
corruption, and in the case of Bulgaria, organised crime (EC, n.d.-b). For each country, the
Commission established a set of benchmarks, against which to measure these countries
progress in each relevant area. These assessments are published in annual progress reports,
compiled on the basis of regular expert missions to both countries and the Commissions
observations (EC, n.d.-j). CVM represents a sui generis tool that has been implemented only once
in a very specific context. No similar mechanism was established following the accession of
Croatia in 2013.51
The EC already had the power to initiate such proceedings with respect to measures adopted after the
coming into force of the Lisbon Treaty (EC 2014g).
49 To our knowledge, there is no academic review of the EU Anticorruption Report. Identified literature
frequently mentions the report and its findings but does not offer an assessment of the tool itself.
50 In addition to developing and managing its own monitoring mechanisms, the EU has also supported a
range of other research projects intended to shed further light on the extent and nature of corruption in the
EU. The largest of these is ANTICORRP, a research project supported by 8m from the ECs Seventh
Framework Programme. Running from 2012 to 2017, the project aims to advance knowledge on how
corruption can be curbed in Europe and elsewhere.
51 According to Pech (2015), political considerations may have influenced the decision not to introduce
CVM for Croatia following the countrys accession, since submitting Croatia to the CVM would have sent
the signal that another country was joining the EU after appearing to lack a functional rule of law system
when its accession treaty was signed. This observation is echoed in Palokaj (2010), who quoted an EU
official as saying we want Croatia to be a good example after the less good examples of Bulgaria and
Romania. For a similar narrative, see Riley (2013).
48
PE 579.319
68
1.3.3. EU Justice Scoreboard

A third monitoring mechanism conducted by the European Commission is the EU Justice
Scoreboard.(EC, n.d.-d) The scoreboard is an annual assessment of Member States justice
systems along three types of parameters that are understood to define the effectiveness of a
given judicial system: a) efficiency, b) quality, and c) independence. In developing the
scorecard, particularly in the areas of efficiency and quality, the EC built primarily on work
done by the CoEs Commission for the Efficiency of Justice (CEPEJ).52 Additional sources used
to develop the Scoreboard include the World Bank, business reports and a network of councils
for the judiciary (who particularly provide input in relation to judicial independence).53 Since
effective justice systems are seen as an integral part of economic development, the results of the
scoreboard feed into the European Semester of Economic Governance, a coordinating
mechanism for economic policy within the EU, where a set of country-specific
recommendations are formulated.54
1.3.4. Other monitoring and research

In addition to the work discussed above, there is a range of other ongoing research projects that
may contribute to the strengthening of monitoring and reporting systems, for instance by
developing new and better indicators of corruption. Relevant initiatives in this field include
ANTICORRP, FRAs project on fundamental rights indicators (FRA, n.d.-b), and work done by
the Government Transparency Institute (Government Transparency Institute, (n.d.)) and the
Anti-Corruption Resource Centre (U4) (U4 Anti-Corruption Resource Centre, n.d.). A
comprehensive resource in this effort is the UNDP 2015 guide on Users Guide to Measuring
Corruption and Anti-corruption (UNDP 2008).
One interviewee also mentioned a compendium of good practice in the field of anticorruption,
prepared by University of Utrecht (2008) and suggested sponsoring the development of a new
iteration, which may form the basis of an updateable catalogue of evidence and effectiveness
data.
1.4. Non-EU law and institutions

The fight against corruption in the EU is also carried out through the efforts of other
international organisations. Three stand out in particular and are discussed in greater detail: the
Council of Europe, the OECD and the UN. The fight against corruption undertaken by each of
these organisations builds on their own legislative framework and has in place accompanying
monitoring systems.
CEPEJ stands for The European Commission for the Efficiency of Justice. For more information, see
(CoE, n.d.-d).
53 Councils for the judiciary are independent bodies that support and manage the administration of justice
in a given country and are present in 20 Member States. In the remaining eight countries, the Justice
Scoreboard works with national supreme courts to collect necessary data.
54 Correspondingly, the need for structural reforms in the area of justice has been listed as an objective in
the EUs Annual Growth Surveys. See, for instance, the 2015 edition (EC 2014b) or the 2016 edition (EC
2015a).
52
PE 579.319
69
1.4.1. The Council of Europe and GRECO

The Council of Europe has produced a suite of instruments intended to contribute to the fight
against corruption in the EU. These include:
-
Criminal Law Convention on Corruption (STE 173, adopted in 1999) and its additional
protocol adopted in 2003.
Civil Law Convention on Corruption (STE 174, adopted in 1999).
Resolution (97) 24 on the twenty guiding principles for the fight against corruption
(1997).
Recommendation No. R (2000) 10 on codes of conduct for public officials.
Recommendation No. R (2003) 4 on common rules against corruption in the funding of

political parties and electoral campaign.
In addition, the Council of Europe hosts the Group of States against Corruption (GRECO),
which oversees members compliance with existing CoE anti-corruption standards. The
overarching objective of GRECOs work is to generate pressure on its Member States to improve
their fight against corruption. The mechanism to do so is through the identification of gaps in
national policies and through the promotion of adequate reforms. GRECOs work is organised
in thematic evaluation rounds (four conducted so far).55. Each round consists of selfassessments conducted by the country being evaluated, monitoring visits by experts
representing other countries, and discussions with relevant country representatives. Based on
the data collected, a final evaluation report is developed, which includes an assessment of the
national compliance with GRECOs standards and a set of recommendations for future action to
improve the level of compliance.56 The uptake of recommendations by individual countries is
monitored in subsequent compliance reports (up to two envisaged per country per evaluation
round).
1.4.2. OECD
In the area of international business, the OECD Convention on Combatting Bribery of Foreign
Officials in International Business Transactions established international standards with respect
to the bribery of foreign public officials (OECD, n.d.-b). Its implementation is monitored by the
OECD Working Group on Bribery, which has developed a peer-review mechanism undertaken
by experts from the member countries. This mechanism has so far consisted of three phases
The four rounds covered the following topics: independence, specialisation and means available to
national bodies engaged in the prevention and fight against corruption, extent and scope of immunities
(Round 1), identification, seizure and confiscation of corruption proceeds, public administration and
corruption (auditing systems; conflicts of interest), prevention of legal persons being used as shields for
corruption, tax and financial legislation to counter corruption, links between corruption, organised crime
and money laundering (Round 2), incriminations provided for in the Criminal Law Convention on
Corruption (ETS 173), its Additional Protocol (ETS 191) and Guiding Principle 2 (GPC 2), transparency of
Party Funding with reference to the Recommendation of the Committee of Ministers to Member States on
common rules against corruption in the funding of political parties and electoral campaigns (Rec (2003) 4)
(Round 3), ethical principles and rules of conduct, conflict of interest, prohibition or restriction of certain
activities, declaration of assets, income, liabilities and interests, enforcement of the rules regarding
conflicts of interest, awareness (Round 4)
56 As per GRECOs Rules of Procedure (CoE, 2011).
55
PE 579.319
70
(OECD, n.d.-a). In the first phase, the focus is on the adequacy of national legislation to
implement the aforementioned Convention. The second phase assesses the extent to which this
legislation is implemented effectively. In the third phase, countries are evaluated on their
enforcement of the Convention and on any outstanding recommendations from previous
phases. A fourth phase is envisaged by the OECD and is currently in the process of being
developed.
In addition to its review mechanism, the OECD has produced a set of soft law measures
intended to strengthen integrity in the public sector. These include, among others, disallowing
the tax deduction of bribes to foreign public officials and measures to strengthen public sector
integrity by providing principles and guidelines for areas such as public procurement and
lobbying (Anagnostou and Psychogiopoulou 2014).
1.4.3. UN
The United Nations adopted in 2003 the aforementioned UN Convention on Corruption (UN
General Assembly Resolution 58/4 of 31 October 2003), which entered into force in 2005 and
includes a set of standards, rules and measures to be applied by signatory parties (UNODC
2004). Appendix F shows that all Member States have ratified the UNCAC. The UNCAC also
includes a review mechanism, intended to assess the progress in its implementation.57 As with
the CoE and OECD reviews discussed above, UNCACs mechanism, managed by the
Implementation Review Group, is conducted on a peer review basis, albeit a lighter touch
review than the others largely conducted on paper. First, the country under evaluation
produces a self-assessment report, which is reviewed by experts from two other participating
countries. At this stage, there may be a supplementary country visit by the reviewers. A final
report is prepared by the reviewers with input and agreement from the country under
evaluation (UNODC 2011).
1.4.4. Other actors

In addition to the mechanisms described above, the G20 has an anti-corruption working group,
established in 2010, which provides another important framework to discuss corruption and
integrity issues at the global (developed country) level. Its current activities are guided by an
Action Plan (G20 Australia 2014a), accompanied by a detailed Implementation Plan (G20
Australia 2014b). The current areas of focus include beneficial ownership transparency, public
sector transparency and integrity, private sector transparency and integrity, bribery,
international cooperation, and vulnerable sectors such as customs, extractives, fisheries and
primary forestry and construction.
In addition to the monitoring and activities mentioned earlier in this chapter, academics,
researchers and NGOs (such as Transparency International and Global Witness, and the
International Consortium of Investigative Journalists) are very active in the field of corruption,
reporting on various aspects of the phenomenon of corruption. While not always of a similar
standing and systematic character, these activities need to be considered as part of the overall
existing framework of corruption monitoring in the EU.
This process is organised in five-year cycles, each focusing on two of the Conventions four substantive
chapters (Transparency International, 2013).
57
PE 579.319
71
2. Effectiveness assessment
This section provides an assessment of the strengths and weaknesses of the measures described
in Chapter 3, Section I i.e. an assessment of effectiveness. Building on discussions in sources
such as the EU ACR (EC 2014h) or the abovementioned systematic review by Hanna et al.
(2010), we consider policies and measures effective if there is evidence of their ability to reduce
the level of corruption (or the perception thereof) or to reduce the opportunity for corruption.
As mentioned in Chapter 1, there is a lack of evidence on the degree of effectiveness of
individual anticorruption measures and almost no evidence of their quantified impact on
corruption levels. This is perhaps not surprising, given the multifaceted character of the
corruption challenge and the importance of contextual factors.58
We have incorporated available information, to the extent possible, in our quantitative analyses
(as presented below) but reiterate that the existing evidence base provides at best an indication
of possible effects and their sizes, let alone any guidance on prioritisation. As with other parts of
this research paper, the assessment of effectiveness in this chapter is based upon a review of
literature and interviews.
2.1. The effectiveness of the EU legislative framework and policy

measures, and of non-EU law
The discussion below is structured in themes that can be understood as barriers to greater
effectiveness of anti-corruption measures. These themes have been identified by the research
team from an analysis of findings from the literature review and interviews. As such, rather
than passing a judgment on individual measures, this section provides a consideration of the
effectiveness across the existing anti-corruption portfolio. This reflects recognition of the multifaceted character of corruption, the importance of context and the dearth of evidence with
individual anti-corruption measures as the unit of analysis.
2.1.1. In some instances, EU and other legislation has not been

formally transposed by Member States
One of the fundamental obstacles to effectiveness of EU action in relation to corruption is the
lack of formal ratification or proper implementation of existing legislation by Member States:
-
The 2003 Framework Decision. In its second implementation report the Commission
noted that the quality of transposition was uneven across Member States and in some
cases incomplete, with some articles having been transposed correctly only by a
minority of Member States (COM(2011)309). Appendix F indicates which Member
States have transposed the Framework Decision.
We note some exceptions to the rule, such as QoGs work on corruption covariates (Rothstein and
Holmberg 2011), similar findings on correlations done by ANTICORRP (Mungiu-Pippidi 2013);
calculations and assumptions in the impact assessment of the EPPO (EC 2011c); and attempts at
quantifying the impact of whistleblower protection by Santoro et al. (2014).
58
PE 579.319
72
The CoEs Conventions: Germany has not ratified the CoEs Criminal Law Convention
(CoE, n.d.-a). 59 Four Member States have not ratified its Additional Protocol.60 The
Civil Law Convention still awaits ratification in six Member States.61
The OECD Anti-Bribery Convention: This has not been ratified by five EU Member
States (none of the non-ratifying Member State is a member of the OECD).62
2.1.2. Some provisions of international legislation are optional or

reserved
A related barrier to effectiveness, mentioned in interviews and in the literature, is where a
legislative instrument is in place but at least some of its provisions are not binding. This limits
the contribution of these instruments to the fight against corruption in EU Member States. This
particularly relates to the UNCAC and the CoE Criminal Law Convention:
-
The UNCAC: Babu (2006) notes that individual provisions of the Convention vary in
the degree to which they are binding. While some obligations are mandatory and
require signatories to undertake legislative action, the implementation of other
provisions remains at the discretion of individual countries.
The CoEs Criminal Law Convention: Anagnostou and Psychogiopoulou (2014) point
out that this is subject to reservations from its Member States (this option is not allowed
under the Civil Convention). This may place some limits on the effectiveness of the
Convention (although the Committee of Ministers has appealed to the Member States to
use their reservation rights as sparingly as possible).
2.1.3. It is too early to determine the effectiveness of recent EU

legislation, for example, in relation to public procurement
As discussed above, a recent legislative development is the adoption of a set of new public
procurement directives. The Sector Directive (2004/17/EC) and the Classical Directive
(2004/18/EC) on public procurement will remain in force until 17 April 2016. Member States
have until April 2016 to transpose the new directives into their national law.63 As such, it is
impossible to assess the effectiveness of the new directives since their implementation is still in
progress.
The enforcement and the effective application of the new rules will be subject to dedicated
monitoring mechanisms64
We note that the recently enacted German law against corruption paved the way towards eventual
ratification of the Convention (Behr 2015).
60Czech Republic, Estonia, Germany, Italy (CoE, n.d.-c).
61 Denmark, Germany, Ireland, Luxembourg, Portugal, United Kingdom (CoE, n.d.-b).
62 Croatia, Cyprus, Lithuania, Malta, Romania (OECD 2014a).
63 One exception is the application of provisions on e-submission, which can be postponed until October
2018.
64 MS will have to report on public procurement activity and they will be obliged to submit to the
Commission every three years a monitoring report covering information on the efficacy and uniform
application of the new directives. See, for instance, Article 83 on Governance of the Directive 2014/24/EU
(EU 2014b).
59
PE 579.319
73
In addition, available literature, and one interviewee, highlighted the role of exclusion of
economic operators (debarment) from public procurement processes on the basis of their
previous criminal convictions. The current rules of debarment stem from provisions in the 2004
public procurement directive and have been transposed into the legislation of all Member
States.65 The relevant 2014 public procurement directive (2014/24/EU), now under
implementation by Member States, introduced two modifications to the rules, making them
somewhat less rigid.66 First, the exclusion period was limited to five years. Second, black-listed
companies are given a chance to terminate the exclusion if they can demonstrate the completion
of a self-cleaning procedure (Manacorda and Grasso 2014; Linklaters 2014). Examples of
actions that may be considered sufficient in this regard include appropriate staff reorganisation
measures or the implementation of an internal audit structure.67 However, Corruption Watch
UK (2015) suggested that these self-cleaning provisions may be difficult to implement in a
coherent manner since procurement officials, being neither compliance specialists nor
enforcement officers, might not have the necessary expertise (Corruption Watch UK 2015). As
such, they may represent a mechanism through which companies will be let off the hook
(Hawley 2015). With respect to the situation at the EU institutions, debarment provisions are
implemented on the basis of the EUs current Financial Regulation (EU 2015b).68 Accordingly,
information on operators excluded by the EU is kept in the Early Detection and Exclusion
System (EDES) (EC, n.d.-f).
2.1.4. There is evidence of variability in prevention, control and

regulation mechanisms within Member States
The area of prevention and control mechanisms is addressed by several CoE documents:
-
The Criminal Law Convention imposes obligations on signatories to adopt necessary

dissuasive mechanisms.
CoEs twenty guiding principles invite individual countries to introduce prevention

and control measures in their legislation. 69
Member States have been provided guidance in the form of CoEs recommendation on
common rules against corruption in the funding of political parties and electoral
campaign (CoE 2003).
Preventative anti-corruption practices are also addressed in UNCAC.70
Prevention measures. When assessing the situation in EU Member States, the EU Anticorruption
Report found variability across individual countries. For example, some countries were found
Article 45 of Directive 2004/18/EC. Crimes constituting the basis for a mandatory debarment include
participation in a criminal organisation, corruption, certain types of fraud, and money laundering. In
addition, the directive lays out conditions for discretionary debarment, such as the state of bankruptcy,
conviction of grave professional misconduct, etc.
66 As Hawley (2015) summed the situation up, the new directives are proportionate and allow for
rehabilitation.
67 See Paragraph 102 of the Directive 2014/24/EC for a full discussion.
68 In particular Articles 106 and 108.
69 See, in particular, Art 1, 4, 5, and 17.
70 See, in particular, Art 5.
65
PE 579.319
74
to have implemented effective prevention measures, while in others this effort remains
fragmented.71
Politicians, political financing and public officials. The ACR found anti-corruption policies appear
to be prominent in the political discourse in the majority of EU Member States.72 In addition,
most Member States have introduced legislation to create higher transparency standards
(compared to previous rules in the country) on political financing. At the same time, elected
bodies and political parties have only infrequently introduced codes of conduct, and even
where these codes do exist, they are not accompanied by effective monitoring and sanction
regimes. The requirements for public officials to disclose their assets were found to have
generally been strengthened across the EU, but substantial differences exist across Member
States with respect to the regulatory regime surrounding conflict of interests.73
Non-regulation of lobbying was a gap in existing practice identified in the country reports and
frequently mentioned by interviewees. Civil society organisations, such as Transparency
International have pushed for more transparency and accountability in relation to lobbying of
the national parliaments (Transparency International 2009). According to the country reports
prepared as part of the Anticorruption report, 22 Member States74 did not have a mandatory list
of registered lobby groups, whether it is business, companies, NGOs, or think tanks. In some of
these countries, however, a register exists on a voluntary basis or the national parliament
introduced a Code of Ethics.75 According to a 2015 TI report based on an examination of 19
Member States, lobbying is regulated by a dedicated law in seven Member States.76
2.1.5. The effectiveness of EU action is dependent upon skills,

capabilities and capacities within Member States
As shown in Appendix F, most, but not all Member States have a national anti-corruption
strategy. Countries differ in the extent to which tackling corruption is a national priority and in
respect of the intuitional arrangements for addressing the problem.
National specialist anti-corruption agencies. As part of their fight against corruption, some
countries have introduced specialised central anti-corruption agencies (e.g. the Federal Bureau
of Anti-Corruption in Austria, the Central Service for Prevention of Corruption in France, and
the Central Anti-Corruption Bureau in Poland), in line with the requirement in the UN
Convention77 and one of the CoEs twenty principles.78 In some instances these agencies have
However, in this context effective is more akin to having some observable effect rather than
demonstrated to be effective through rigorous policy evaluations.
72 COM((2014)38
73 COM((2014)38
74 The countries are the following: Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia,
Finland, France, Germany, Greece, Italy, Latvia, Luxembourg, Malta, the Netherlands, Portugal, Slovakia,
Spain and Sweden.
75 For instance, Germany and the Netherlands introduced a voluntary register, but this is not regulated by
law. Ireland introduced a Lobbying Regulation bill which made it mandatory to register lobbying
activities and made the information available to the public. However, it fell short in its focus on
responsibilities of lobbyists and, to a lesser extent, with respect to those of public officials (EC 2014d).
76 These are Austria, France, Ireland, Lithuania, Poland, Slovenia and the United Kingdom. TI (2015)
Lobbying in Europe: Hidden influence, privileged access (Transparency International 2015b).
77 Article 6.
78 Article 7.
71
PE 579.319
75
led to notable and sustainable results, while in other cases the results were more mixed,
suggesting that the establishment of dedicated agencies is not a silver bullet in the fight against
corruption (COM(2014)38).79 An interviewee observed that anticorruption agencies might
represent a good example of the tension between the formal existence of anti-corruption tools
and their effectiveness, in particular in instances where control over anti-corruption agencies
has been politicised and their work diverted to further domestic political goals. The interviewee
added that this risk was mainly relevant for current and recent candidate countries.
Challenges related to the independence of law enforcement and prosecution agencies charged with
responding to corruption have also been raised in some Member States, though it is not always
obvious what would constitute sufficient resources compared with other social objectives (Lord
and Levi 2015). In some Member States, anti-corruption efforts are reported to be hampered by
insufficient capacity and/or determination of the judicial system to address some corruption
cases.80 As with other topics discussed in this section, obligations and recommendations in this
area have also been formulated in existing international instruments, such as CoEs twenty
guiding principles (Article 3) and the UNCAC (Chapter III).
Insufficient administrative capacity, and the associated gaps in the implementation of anticorruption measures, at least in the context of some newer Member States, may be a reflection
of the challenges faced in improving their administrative capabilities. For instance, as OECDs
2009 SIGMA (Support for Improvement in Governance and Management) report noted, civil
service reforms undertaken in CEE countries in preparation for their accession to the EU have
been stalled or reversed in the majority of the countries after they joined the EU (Meyer-Sahling
2009). This is in line with an observation made by several interviewees that the EU loses a
substantial amount of leverage vis--vis individual countries at the moment of a countrys
accession. The same sentiment was expressed by Commissioner Viviane Reding, who, in the
context of rule of law stressed the existence of a Copenhagen dilemma:
We face a Copenhagen dilemma. We are very strict on the Copenhagen criteria, notably on the rule of law
in the accession process of a new Member State but, once this Member State has joined the European
Union, we appear not to have any instrument to see whether the rule of law and the independence of the
judiciary still command respect.(Council of the EU 2012)81
This observation is also made by ANTICORRP researchers, who observed that countries with special
anti-corruption agencies do not appear to fare significantly better than countries which address corruption
via their normal legal system (Mungiu-Pippidi 2013). However, there may be some value in having a
dedicated anti-corruption body even if it is not particularly effective, such as symbolic reassurance and
bolstering of the states legitimacy in the fight against corruption.
80 The abovementioned CEPEJ reports and the EU Justice Scoreboard are comprehensive sources of
broadly comparable data on relevant indicators in this area.
81 However, as a possible response to this dilemma, please see discussion of measures to protect rule of
law available to the EU in section 3-III-1.
79
PE 579.319
76
2.1.6. The effectiveness of the fight against corruption might be

reduced by gaps in EU law
Gaps in current EU law were identified in the available literature and by interviewees.
Definition of public official

The EU ACR noted that there is no harmonised definition of public official at the EU level that
would include people holding an elective office. One interviewee suggested that the underlying
barrier is different definitions and understanding of a public official across Member State.82 These
range from a narrow definition of people who only work in the government to a second
approach including everyone starting at local level up to the government and works in a public
administration, as well as MPs and MEPs. According to the interviewee, the consequence of this
inability to come up with a unified definition is the abovementioned absence of a
comprehensive EU instrument on corruption in the public sector (to complement the existing
framework decision covering the private sector). 83
As further discussed in Section III, below, a definition of a public official is included in the draft
text for a new Fraud Directive (also referred to as PIF-Directive).
Protection for whistleblowers

Another gap in EU legislation highlighted by interviewees is the absence of an EU instrument
on the protection of whistleblowers. Protection for such individuals is internationally
recognised as essential to fight corruption84 and is already featured in a number of international
instruments (see Box 6).
Box 6: Calls for the protection of whistleblowers from the EU, UN, OECD and CoE
The UNCAC requests that Member States consider whistleblower protections in their
domestic legal systems.85
A similar recommendation to introduce whistleblower protection measures is made in
OECDs 1997 Convention on Combatting Bribery of Foreign Public Officials and in CoEs
Twenty Guiding Principles for the Fight Against Corruption and Civil Law Convention on
Corruption.86
A recent case demonstrated the persistent lacking of a global agreement on the definition of public
officials. Germany did not ratify UNCAC due to its parliamentarians reluctance to relinquish rights
under legislation on bribery of members of parliament until September 2014. Under the UN convention,
Members of Parliament are considered public officials (Transparency International 2014a).
83 In this context, however, we reiterate the lack of legal basis for an EU-led establishment of common
corruption standards across Member States.
84 For instance, among the first instruments on whistleblowing were the OECD Recommendation on
Improving Ethical Conduct in the Public Service in 1998 including the Principles for Managing Ethics in
the Public Service (OECD, 1998) and the OECD Recommendation on Guidelines for Managing conflict of
Interest in the Public service of 2003 (OECD 2003).
85 UNCAC, Article 33
86 Article 9 of the Convention states that Each Party shall provide in its internal law for appropriate
protection against any unjustified sanction for employees who have reasonable grounds to suspect
corruption and who report in good faith their suspicion to responsible persons or authorities. However, as
Stephenson and Levi (2012) noted, this is a rather loose provision, since it remains ambiguous with
82
PE 579.319
77
The Parliamentary Assembly of the Council of Europe (PACE) asked its Member States to
review their whistleblowing legislation (Resolution 1729 (2010)) and requested that the
Committee of Ministers (CoM) prepare a set of guidelines for the protection of
whistleblowers (Recommendation 1916 (2010)). Following a feasibility study on a
possible new legal document (Stephenson & Levi 2012), the CoM issued a
Recommendation on the Protection of Whistleblowers (CoE 2014b).
Simultaneously with CoEs activities, the EU has also focused on whistleblower
protections. For instance, in its communication to the Council, the EP and the EESC on a
comprehensive policy against corruption (EC 2003), the Commission called on Member
States to introduce measures for the protection of whistleblowers. This recommendation
was reiterated in a recent LIBE report on the situation of fundamental rights in the EU (EP
2014c).
Despite the existence of these instruments, whistleblower protection in EU Member States

remains patchy.87 While whistleblower legislation exists in some Member State, this type of
protection is not widely available across the EU.88
Insufficient provisions for the protection of whistleblowers have also been noted with respect to
corruption in European institutions. A 2011 study on the effectiveness of whistleblowers
conducted by PwC and i-Force (2011) found that current whistleblowing rules were not
effective for two reasons. First, the rules were unclear and too narrow in scope and second, their
implementation had suffered from a series of deficiencies.89 These observations are in line with
the findings in Transparency Internationals Integrity System Report, which observed that, with
the exception of the European Commission, no EU institution has in place a set of guidelines
and procedures to protect whistleblowers, despite the existence of EU Staff Regulations
requiring such procedures be put in place (Transparency International 2015a).90 As the TI
Integrity System Report noted, this is particularly noteworthy in the case of review and
oversight bodies such as OLAF and ECA (Transparency International 2014b).
The TI report was followed by an inquiry led by the European Ombudsman, which found that
only two of the nine institutions covered91 adopted rules of the kind required (European
respect to what necessary means. Still, the authors pointed out that the Convention had influenced ECHR
case law.
87 For an overview of existing whistleblower legislation at the national level, see (European Univerity
Institute, n.d.).
88 The need for an effective whistleblower protection system was recently acknowledged by some MEPs in
the aftermath of the LuxLeaks scandal, in which the principal whistleblower was indicted for disclosing
information on tax avoidance schemes in Luxembourg (EP 2015e).
89 These included a lack of integrated organisational approaches, the absence of an independent helpdesk,
insufficient independence of OLAF (or at least perception thereof), and a lack of a working system for
tracing disclosures.
90 In addition, an earlier 2006 study found that the scope of Staff Regulations (notably its Articles 22a and
22b) covered only a part of what may typically fall under the category of whistleblowing. The report also
found that the regulations are of limited effect in promoting desirable behaviour both on the management
side as well as on the staff side due to the fact that they only promise that EU Institutions will not react
negatively to a report made by a staff member (Rohde-Liebenau 2006).
91 These were the European Parliament, the European Commission, the Council of the European Union,
the Court of Justice of the European Union, the European Court of Auditors, the European External Action
PE 579.319
78
Ombudsman 2014). The lack of systematic whistleblower protection was also noted in the EPs
draft report on Transparency, accountability and integrity in the EU institutions92 and by
several interviewees, who called for stronger action by the EU in the area (EP 2015k).
Protection of whistleblowers was also mentioned by interviewees as a desirable step in
connection with public procurement. The same observation was made by Anagnostou and
Psychogiopoulou (2014), who noted that local and regional governments in CESE countries are
particularly problematic in this respect, as they appear to be the weakest performers in the field
of whistleblowers protection. Even when mechanisms are in place, citizens do not seem to trust
them.
In the view of the interviewees, an EU-level instrument may be a suitable way to address this
situation.93 The absence of an EU-wide instrument on whistleblowing is also noted by Santoro
et al. (2014), who as part of their work for the Restarting the Future initiative called for the
adoption of an EU directive on whistleblowing, followed by the establishment of a European
Authority for Whistle-blowing.94 This is discussed further in Chapter 3 Section III.
2.2. The effectiveness of existing monitoring mechanisms

2.2.1. EU Anticorruption Report
At the time of writing there has not been a follow-up assessment of the extent to which Member
States have acted on these recommendations. Some follow-up on progress is envisioned for the
next iteration of the report, though not based on formal review mechanisms such as those put in
place by GRECO or UNCAC. This section is based on the views of interviewees and expressed
in the literature.
Elements perceived to have been effective

The ACR has raised the profile of the fight against corruption: Interviewed experts
unanimously recognised the importance of ACR and its high profile as a flagship monitoring
project of the EU. One interviewee saw the report as a demonstration of the ECs effort to
elevate the profile of the fight against corruption and help integrate it into other EU policies.
Another interviewee felt that it would take some time for the report to gain visibility because it
is a relatively new process.
Service, the European Economic and Social Committee, the Committee of the Regions and the European
Data Protection Supervisor.
92 EP (2015) Transparency, accountability and integrity in the EU institutions. Committee on Constitutional
Affairs
Draft
Report.
(2015/2041(INI)).
https://polcms.secure.europarl.europa.eu/cmsdata/upload/adc420de-e7a8-4bae-9e4f4acedf06fb50/AFCO_PR(2015)567666_EN.pdf. (As of 23 February 2016)
93 As another example of a gap in EU legislation, another interviewee noted that the 2003 Framework
Decision does not cover civil confiscation, a procedure that has been introduced in a number of MS. This
issue is covered in the cost of non-Europe paper on organised crime.
94 Restarting the Future is a public campaign led by civil society organisations and enjoys the support of 69
MEPs from six political groups. For more information see (Restarting The Future, n.d.).
PE 579.319
79
The ACR brings together a range of indicators relevant to corruption: The reports ability to
draw on disparate data sources and bring together available relevant data was valued by
interviewees.
The ACR produces tailored country reports: Interviewees noted the fact that the report
identifies the most problematic areas and produces recommendations for individual Member
States on how to proceed. One interviewee highlighted the fact that the ACR reflects the
varying needs of individual countries by producing country-specific reports that focus on the
most pertinent issues for the country in question. The expert noted that it would have been
easier to produce a uniform matrix-style information product, but felt that such a product
would have been of very limited use.95 While this approach to organising the document was
difficult to communicate to individual Member States, it was ultimately accepted and the fact
that this discussion is now settled has laid the foundation for a regular useful product going
forwards.96
The ACR is complemented by the anti-corruption experience-sharing programme: An
additional important product of the ACR, highlighted by two interviewees, was the
establishment of the anti-corruption experience-sharing programme, intended to serve as a
platform for interested parties and stakeholders to discuss how to best address challenges
identified in the ACR (EC, n.d.-a). So far, there have been three workshops organised in the
framework of the programme, covering the areas of asset declaration, whistleblowing and
healthcare corruption.97
Elements perceived to be barriers or limitations

The ACR includes little new data: One interviewee pointed out that since the ACR draws
heavily on work conducted by others (though it does include some original data collection), it
may face an ongoing challenge in demonstrating its added value and usefulness. This is related
to the point about risks of duplication mentioned before in section 3-II-1. However, the
interviewee felt that the biggest added value of the ACR may be its profile-raising effects
mentioned above, i.e. putting anti-corruption policy on the agenda and creating opportunities
for exchanging ideas and policy integration. Even if the EC is criticised for compiling existing
evidence rather than conducting new analysis, it acts as a focal point for reform.
There is no formal assessment procedure: While some follow-up on country-specific
recommendations is envisaged for its next iteration, no formal assessment procedure, for
instance akin to GRECOs evaluations, has been established within the framework of the EU
ACR. With respect to recommendations, a TI assessment of the ACR lamented the lack of
At the same time, comparisons can be made with GRECOs evaluations, which are standardised and
follow a predetermined agenda. One interviewee noted the GRECOs systematic approach is well received
by the Member States. A standardised structure may be one way to avoid pushback from countries under
evaluation, who may be concerned about differential treatment.
96 Admittedly, the format of the report can be understood to be a barrier in so far as it places limits on
comparability across Member States.
97 The event has the benefit of bringing together a diverse range of stakeholders with different perspectives
and experiences on the issue of corruption, reflecting its horizontal and cross-cutting character. Using the
workshop on whistleblowing as an example, the sessions brought together anti-corruption experts, and
representatives of NGOs and public authorities, including criminal justice stakeholders (police officers,
prosecutors) and relevant national ministries (labour, finance, justice and interior). Outputs from each
workshop are available online (EC, n.d.-a).
95
PE 579.319
80
specific recommendations on the protection of whistleblowers, access to information and

lobbying. In addition, TI noted the absence of attention to the cross-border dimension of
corruption (Dolan 2014).
Corruption at the EU level does not fall within the scope of the ACR.98 Therefore, at least
from the perspective of the EU and its report, monitoring and assessment procedures with
respect to corruption at the EU level are weaker than their counterparts pertaining to the
situation in the Member States. In an effort to remedy this gap, Transparency International
prepared a review of corruption and integrity risks in the EU institutions and pertaining to the
EUs budget in its 2014 EU Integrity System Report (Transparency International 2014b). This
report contains an assessment of ten EU institutions and agencies (EP, European Council,
Council of the EU, EC, CJEU, ECA, OLAF, Europol and Eurojust, and Ombudsman) and
represents, to our knowledge, the most comprehensive independent report on corruption at the
EU level to date. The EU and its institutions are subject to external review through its
membership in UNCAC; however, such a review has not been completed yet. A similar
scrutiny would also be applied to the EU and its institutions should it accede to GRECO (as
discussed in Section III of this chapter).
2.2.2. CVM
The CVM has been an important lever to push for reform: Available academic literature
widely acknowledges the role of the CVM, while taking note of its limitations. For instance,
Vachudova and Spendzharova (2012) found the CVM to be indispensable in pressuring the
Bulgarian and Romanian governments to adopt and implement key institutional reforms. The
Romanian Institute for Public Policy (2010) credited EU monitoring mechanisms, starting with
pre-accession progress reports, with establishing and elevating the fight against corruption on
the domestic political agenda.
CVM is a capacity-building tool: One interviewee noted the CVM can be seen as a capacity
building tool. For instance, with the help of annual progress reports, in the view of the
interviewee, the judicial system (mainly judges and prosecutors) became more self-confident
about their rights to become a driver of the fight against corruption. This capacity building and
mobilisation aspect of CVM has been examined also in relation to civil society actors in Bulgaria
and Romania. Dimitrova and Bugozany (2014) found that the CVM served as a point of
reference for the activities of civil society organisations and the media. This contribution has
also been recognised by Ivanova (2009) and Vachudova and Spendzharova (2012).
CVM is integrated in wider reforms: Interviewees felt that the CVM has had a much bigger
impact than the ACR, which might be expected given the difference in the size of the projects
and how long they have been in place. One possible contributing factor to the achievements of
the CVM in the area of corruption, stressed by one interviewee, was the fact that it has been
integrated into a wider reform process (governance, civil society formation).
Although we note that a separate chapter covering the EU level was originally envisaged and drafted
(EC 2013d).
98
PE 579.319
81
CVM receives support from domestic authorities: Interviewee felt that the CVM has had very
good buy-in from official authorities who identified themselves with the reform and wanted to
be seen in Brussels as doing a good job. Similarly, authors such as Vachudova and
Spendzharova (2012) and Dimitrova (2014) observed that domestic political considerations were
often an important enabler of progress and a source of incentives, in particular for political
parties and representatives who presented themselves as pro-European.99
CVM reports are evidenced-based and detailed: The CVM annual reports have been described
as highly technical, granular and well documented (Vachudova & Spendzharova 2012). One
interviewee observed that the reports did not elicit any objections and were accepted by
national authorities because they were well-evidenced and argued. Similarly, Pech (2015)
commended the mechanism for establishing well-defined and easily observable progress
indicators.
On the whole, interviewees agreed that the two countries covered by the CVM have made
notable progress while covered under the mechanism and one interviewee felt strongly that the
CVM has had a very positive effect on the situation in both countries, even if to a different
extent.100
Elements perceived to be barriers of limitations

Evidence of actual change is mixed: Dimitrova (2015), while acknowledging the CVMs
achievements in terms of improving government actions and laws, notes that with respect to
actual behavioural change, the CVMs record is rather mixed and remains subject to debate. At
the opposite end of the assessment spectrum, studies (Dimitrov et al. 2014; Dimitrov,
Haralampiev & Georgieva 2014) found the CVM to be largely ineffective. This conclusion was
based on an observation that while the EU monitoring focused excessively on formal legislative
change, it failed to generate change at the societal level.101
CVM might not be a transferable anti-corruption mechanism: Interviewees commented on the
effectiveness of the CVM with the caveat that it is a very specific tool intended for a very
specific context (and therefore might not be a transferable approach to other Member States).
CVM has been technical and costly to implement: One interviewee stressed that the CVM has
been a very costly mechanism and has required many skilled individuals to process large
amounts of data. This is not necessarily a criticism of the mechanism itself but rather an
acknowledgement that its replication or adaptation would require substantial investment.
However, taking a forward looking perspective, whether this desire to impress will be sustained to the
extent that there is more scepticism about the EU remains to be seen.
100 At the same time, interviewees pointed out that the very existence of the mechanism can be seen as a
failure since it recognises the fact that Bulgaria and Romania were admitted to the EU prematurely. In the
view of one interviewee, this is particularly notable since the prospect of admission is one of the strongest
levers the EU has vis--vis corruption in third countries. Once new countries are admitted, the strength of
this lever is much weakened. Vachudova and Spendzharova (2012) also note the difference in the leverage
available to the EU pre- and post-accession. However, this is not meant as a criticism of the tool itself but
rather of the need for such tool in the first place.
101 In a similar vein, for an at best mixed view of the CVMs effectiveness in fostering democratic
institutions and rule of law, see Sedelmeier (2014). Still, the author notes some notable achievements and
contributions of the tool, such as work to stop the referendum to impeach President Basescu in Romania in
2012.
99
PE 579.319
82
A risk that CVM might reduce need for internal controls: The same expert who noted the cost
of the CVM also wondered whether the mechanism inhibited, to some extent, the development
of effective internal control mechanisms by the two participating countries. These Member
States may have felt less need to work on further developing their own control systems given
the knowledge that the CVM would perform this function.
CVM does not have strong sanctions: An interviewee pointed out that the sanctions in the
mechanism were weak, thereby limiting enforcement options on the part of the EU. A similar
observation about the limited enforcement options has been made by several academic authors.
Carrera et al. (2013) noted that the CVM represents a soft-policy non-legally binding tool and
Alegre et al. (2009) found the CVM a relatively weak implement with limited powers to press
states into reform. At the same time, studies such as those conducted by Dimitrova (2015) and
Gateva (2010) observed that the CVM developed over time elements of conditionality, which
could be understood as a form of hard sanctions. Two modalities of this conditionality stand
out in particular:
-
Following a report that found substantial deficiencies in the management of EU funds,

the EU decided to freeze several funding streams available to Bulgaria (EC 2008b).102
In 2011 the two countries progress (or lack thereof) in the fight against corruption and
organised crime, as reported in CVM reports, was linked with their Schengen accession,
preventing the countries from joining the border-free agreement (Vachudova and
Shpendzharova 2012).
2.2.3. Justice scoreboard

Data in the scoreboard is independent and validated: In the opinion of one interviewee, one of
the strengths of the scoreboard is the fact that the data informing the Scoreboard is largely
validated through prior data collection exercises and therefore is not subject to any challenges at
the point at which the scoreboard is published. Additionally, the preparation of the tool is not
reliant on Member State political authorities and relies instead on alternative sources, including
a network of independent experts at the country-level.
The Scoreboard is integrated into the European Semester of Economic Governance: Another
feature of the scoreboard highlighted by an interviewee is its integration into the European
Semester of Economic Governance. That way, the scoreboard enjoys access to a mechanism
through which country-specific recommendations are made and follow-up assessments
conducted, thereby creating a cycle of continued progress monitoring.103 This coordination
takes a procedural form as well the publication of the scoreboards findings is scheduled to
coincide with the release of country specific reports in the European Semester.
The funds were unfrozen following a positive 2010 CVM report. Dimitrova (2015) points out that this
option is available even in contexts without the CVM as the Commission has resorted to this step with
Hungary as well.
103 However, the link between country-specific recomomendations and scoreboard indicators is not direct.
Recommendations are formulated on the basis of multiple input sources.
102
PE 579.319
83

The Scoreboard does not have enforcement mechanisms: Butler (2012) pointed out that the
scoreboard does not have, by itself, any enforcement or sanctions provisions. Through its
integration in the European Semester, its findings are reflected in the Semesters countryspecific recommendations; however, Member States cannot be sanctioned for nonimplementation of any reform recommendations.
The Scoreboard has a limited scope: Currently, the Justice Scoreboard is limited in its scope in
that it covers only civil, commercial and administrative justice. The Commission would like to
include (and this is encouraged by others, including the European Parliament104) to include
criminal law as well. However, this is currently not considered feasible by the EC due to the
lack of available comparative data that would enable the construction of the scoreboard in the
criminal domain.105
2.2.4. GRECO
With respect to non-EU monitoring systems, GRECO evaluations elicited varied comments
from interviewees and in reviewed literature.

GRECOs approach is systematic: One interviewee praised GRECOs systematic and thematic
approach to their evaluations, as opposed to what they viewed as a rather subjective selection of
topics for review identified by the EU in its ACR.
GRECO can mobilise soft enforcement levers: One interviewee offered a different assessment
and felt that GRECO had been successful in following up on its recommendations, although the
expert acknowledged that the system is based on political pressure rather than a hard sanction
regime.106 A special procedure is foreseen for addressing situations where a countrys response
to GRECOs recommendations is found unsatisfactory (Rule 32) (CoE, 2011; CoE, n.d.-f). This
procedure entails an intensification of communication between GRECO and relevant authorities
from the country in question, including requests for clarifications and follow-up from the
GRECO president to national delegations, its Ministry of Foreign Affairs and any other
pertinent high-level representative. In addition, GRECO also has the ability to send a mission to
the country and ask for explanation. According to the interviewee, this procedure has worked
very well in the past when applied to Austria and Luxembourg and is well received by the
Member States.
GRECO, along with the UN and OECD, benefits from open channels of communication
among entities that undertake similar activities. As confirmed by representatives of various
monitoring mechanisms, GRECO and organisations such as the OECD and the UN consult
regularly on their activities, attend each others events and offer expertise assistance. This has
See, for instance, the Parliaments Resolution of 12 March 2014 on evaluation of justice in relation to
criminal justice and the rule of law (EP 2014b).
105 One interviewee stressed this data availability issue does not necessarily cover all Member States. In
fact, some countries already report criminal law data in a sufficient quality.
106 According to the interviewee, it is also important to keep in mind another fundamental difference
between the CoE and the EU. Unlike the EU, the CoE does not provide any substantial financial support to
its Member States and thus does not have at its disposal what may amount to a very effective lever.
104
PE 579.319
84
not necessarily translated into a formalised instrument of coordination but has contributed to
the effectiveness of their activities.
Questions about the validity of GRECOs evaluations: Wolf (2010) observed that some
countries scoring well in GRECO evaluations nevertheless fare badly in other corruption
indicators.107
Lack of an effective hard enforcement mechanism: Another interviewee noted that a barrier
to greater effectiveness of CoEs conventions and its monitoring mechanisms was the lack of an
effective enforcement mechanism. In the interviewees opinion, this makes GRECOs work
relatively toothless. This is in contrast to the point made by another interviewee, explained
above, that the soft enforcement mechanisms employed by GRECO were effective.
Effectiveness of GRECO is limited by resources: As with other monitoring mechanisms, the
work of GRECO is subject to capacity constraints both in evaluation and in country
implementation. This has been echoed by one interviewee, who felt that GRECO would benefit
from additional resources that could translate into greater monitoring and reporting capacity on
the ground in individual Member States (of course you get different results if you have 20 on a
single country than having 20 persons for 49 countries).108
2.3. The effectiveness of EU institutions with relevant powers in

addressing corruption at the EU level
The TI report pointed out that OLAF (as well as the ECA and the Parliament) exercise their
oversight roles in an active manner: The EP is issuing an increasing number of requests for
information to other institutions.
There are a number of anticorruption and transparency practices already in place to tackle
corruption within EU institutions: Overall, TIs report on the European Integrity System found
that there is good foundation to tackle corruption laid by policies and rules adopted to tackle
fraud and corruption within the EU (Transparency International 2014b). In particular, the
authors laud the existence of a good number of anticorruption and transparency practices
already in place, such as measures to protect EU institutions and their staff from undue
influence,
and
the
existing
system
enabling
the
investigation
of
alleged
fraud/corruption/maladministration. With respect to tools for engaged citizens, the EU has
provisions in place for the public to ask for information, submit complaints and request review
of its policy decisions.
We note this observation is related to the issue of formal compliance, listed as one of the limitations
mentioned in the introduction chapter to this paper.
108 A similar capacity issue may be faced by the Implementation Review Group of the UNCAC. In 2013, TI
published a progress report on the UNCAC, which found that the review process was substantially behind
schedule, perhaps partly due to its large scope. The report also noted that the Convention included no
clear procedure that would enforce compliance if state parties were found in breach of their obligations.
107
PE 579.319
85
However, the TI report also noted that corruption risks persist at the EU level, largely due to
issues such as poor practice, lack of political leadership, failure to allocate sufficient staff and
lack of clarity about to who the rules apply (Transparency International 2014b).

The perceived independence of OLAF might be undermined due its position within the
European Commission. TI notes that the independence of OLAF, or at least the perception
thereof, may be somewhat undermined. This observation was not necessarily in line with views
expressed by experts interviewed.109 None of the interviewees commented on OLAFs
institutional position as part of the EC as a potential issue and one interviewee explicitly did not
find this set up problematic.110 Still, however independently OLAF actually operates,
perceptions may not correspond to that reality and therefore remain a problem for legitimacy.
OLAF and other institutions rely on Member States to initiate prosecutions regarding the use
of EU funds. Since there are no judicial and prosecutorial powers at the EU level (OJEU does
not have the power to judge EU-level corruption cases) (Transparency International 2014b), the
current system is based on national authorities prosecuting cases based on the 1995 convention
on the protection of EUs financial interests mentioned above. All investigations falling under
the scope of the convention take place either based on Member State initiatives or based on
OLAF investigation. Since OLAF has no competences with respect to criminal investigations, its
activity is limited to making recommendations to individual Member States about whether to
prosecute a given case or not. Importantly, as several interviewees noted and as reported in
annual OLAF reports, the current rate of convictions following OLAFs recommendations is
very low, in some Member States as low as 0 per cent (i.e. no OLAF recommendations lead to
convictions) (OLAF 2011), although there is substantial variability in the conviction rate across
individual Member States.111 . OLAF is currently able to provide recommendations only and has
no power to compel individual Member States to act on these recommendations (Transparency
International 2014b).112 Where Member States decide not to act this may be so for a variety of
reasons, from corruption to inadequate resources afforded to serious fraud and corruption,
whether EU-related or national:113
-
In some instances, Member States have little interest in taking cases forward, for
instance due to the lack of a sense of ownership and due to the difficulty with
identifying the appropriate jurisdiction (particularly for cross-border cases).
A similar comment about the (perceived) lack of OLAFs independence was made in an impact
assessment of the proposed European Public Prosecutors Office (EC 2011b).
110 By contrast, one interviewee felt that, while acknowledging the existence of its supervisory committee,
OLAFs oversight is not as strong as that of its national counterparts.
111 A similar variability (with some countries recording 0 per cent) exists for indictment rates (OLAF 2014).
An interim report by the LIBE committee on the establishment of the EPPO stated that the average rate of
indictment based on OLAFs recommendations between 2006 and 2013 was 31 per cent (EP 2015a).
112 This was echoed by one interviewee who stated that OLAF had no real power. The same observation
applies to the EUs law enforcement agencies, which, as pointed out in the Initial Appraisal of the Impact
Assessment, have been increasingly active but lack necessary powers (Davies 2013). While Eurojust can
ask Member States to initiate an investigation, it has no control of the follow-up. Similarly, Europol
provides support to national authorities but cannot direct national investigations or influence any followup work stemming from its analyses and outputs (EC 2013a).
113 See e.g. Doig and Levi (2013).
109
PE 579.319
86
Authorities may also have little appetite for prosecuting cases with a potential for
conflicts of interest if it is related national authorities who are implementing the EU
funds (EC 2013a).
Lack of resources may also play a role in some cases, coupled with insufficient and
frequently ineffective mechanisms for international cooperation and information
exchange.114
In addition, one interviewee pointed out that in instances where Member States do act
on OLAFs recommendations, their authorities repeat the investigation. This is not only
costly in terms of resources, but also takes considerable time, which may cause some
cases to be time-barred where there is a statute of limitations.
In response to the challenges presented above, the Commission has put forward a proposal to
set up the European Public Prosecutors Office (COM(2013)0534), as envisaged in the Lisbon
Treaty,115 which is currently undergoing consideration. This is discussed further in Section III of
this chapter.
Lack of transparency in relation to law making and lobbying within EU institutions: One
area of deficiency identified by TI concerns transparency in law making and the related issue of
lobbying. Substantial amounts of information and documentation pertaining to the work of EU
institutions are routinely made available.116 However, these publications do not cover all
negotiations and meetings, leaving part of the decision process out of public scrutiny. There is
also no requirement on EU representatives to report on and disclose their meetings with
lobbyists, or on input lobbyists may have had into legislative documents. This concern about
rules governing interactions with lobbyists and vested interests was echoed by several
interviewees, one of whom pointed out that the existing register of such interactions at the EU
level remains optional.
Room for improvement in the rules on conflict of interest by EU officials: With respect to
conflict of interest, the TI found that current rules in place represent a good basis to prevent
corrupt behaviour by EU staff but noted their complexity, creating potential for confusion. With
respect to MEPs and other senior EU figures, gaps exist in the existing control mechanisms,
such as insufficient verification of asset declarations or limited independence of ethics
committees in EU institutions, which are frequently composed of former or current members of
the very institutions they are supposed to oversee.
For instance, a study on the impact of the different policy options to protect the financial interests of the
Union by means of criminal law, conducted by Ecorys and referred to in the EPPO Impact Assessment (EC
2013a), noted that mechanisms such as requests for mutual legal assistance or join investigation teams
often do not function well, due to practical difficulties such as language problems and differences in legal
systems, and are rather lengthy.
115 Art 86 of the Treaty provides for the possibility of setting up a European Public Prosecutors Office
(EPPO) for investigating, prosecuting and bringing to judgment [...] the perpetrators of, and accomplices
in, offences against the Union's financial interests.
116 The Ombudsman has played a role in further improving the transparency of EU institutions.
114
PE 579.319
87
3. Is there potential for action at EU level that might lead to

added value to the challenges identified?
This section presents suggestions on how to address gaps and barriers in existing EU actions,
including legislation and its implementation which hinder effective combat against corruption
in the EU, set out in Section 3-II.
Possible actions that might lead to added value were sought in the available literature and
during interviews. Table 17 shows the potential actions that were identified.
The right hand column of the table lists soft measures. A detailed discussion of these is outside
the scope of this paper as the remit of this paper is limited to EU measures and action and
focuses on modifications or additions to existing legislative instruments. However, several
interviewees suggested that, as a practical consequence of the limited EU competencies in the
area and the need to work with Member States, some of the most actionable and feasible
options for policy action may revolve around soft measures. These measures would be
intended to help generate positive change in individual Member State policy and practice and
may include capacity building, encouragement of political will, etc. In this report, we pay
attention to instances where soft law and soft measures may be appropriate, but their
detailed discussion remains out of scope of this report.
Each of the possible actions in the left hand column is discussed in turn below.
The EU might not be the best actor

When considering future policy options, several interviewees stressed that it should not be
automatically assumed that the EU is the best entity in a position to act. The EU is not the only
actor in the fight against corruption in the EU - GRECO, OECD and the UN play important
roles. As exemplified by the data synthesis supporting the EU Anticorruption Report, there is a
good degree of coordination across these actors and EU activities build on work undertaken by
these external actors. This existence of a multitude of actors, albeit with somewhat differing
remits and activities, means that careful thought should be given to what the EU institutions
should focus on to avoid duplication of effort. As Doig and Levi (2009) note in their review of
inter-agency cooperation in the UK, so much time may be spent on liaison that there is little
resource left to act (Doig & Levi 2009).
Table 17: Potential actions at EU level that might lead to added value to the challenges
identified
Hard measures, within EU competency,

within remint of LIBE committee
-
Make use of infringement proceedings

against Member States who have not
implemented EU law in relation to the fight
against corruption
PE 579.319
Soft measures mentioned in literature and

interviews, measures falling outside of EU
competence, and outside remit of LIBE Committee
(outside the scope of this research paper)
- Education and capacity building of national and
local public authorities
- Further development of mechanisms for sharing
good practice
88
Hard measures, within EU competency,

within remint of LIBE committee
-
Support new legislation to harmonise

protection for whistleblowing within
Member State and/or provide protections
to whistleblowers within European
Institutions
Support new legislation to define a public
official
Make improvements to the ACR
Extend the Cooperation and Verification
Mechanism to other Member States
Make improvements to the EU Justice
Scoreboard
Take steps for the EU to accede to GRECO
Establish a European Public Prosecutors
Office
Soft measures mentioned in literature and

interviews, measures falling outside of EU
competence, and outside remit of LIBE Committee
(outside the scope of this research paper)
- Leveraging enlargement vis-a-vis third countries
- Continuing mainstreaming corruption into other
policy areas
- Introducing greater conditionality of access to EU
funds
- Monitor implementation of legislation under
transposition, primarily procurement directives
3.1. Make use of infringement proceedings against Member States

who have not implemented EU law in relation to the fight against
corruption
This action could address the following gaps and barriers in existing EU actions, including
legislation and its implementation:
-
In some instances EU and other legislation has not been formally transposed by Member
States.
There is evidence of variability in prevention, control and regulation mechanisms within
Member States.
The effectiveness of EU action is dependent upon skills, capabilities and capacities within
Member States.
3.1.1. What would this option involve?

The EU has the possibility to initiate infringement proceedings against individual Member State
with respect to the implementation of EU law at the Member State level. Since December 2014,
this option covers Union acts in the area of police and judicial cooperation in criminal matters
(this equalisation with other policy areas was explicitly recognised in the 2014 Annual Report
on monitoring the application of EU law) (EC 2015d).
In practical terms, infringement procedures follow a well-established process (in accordance
with Article 258 of TFEU) (EC, n.d.-e). The procedure can be linked (under Article 260 of TFEU)
to financial sanctions, following a clear set of guidelines for setting their amount, thus linking
PE 579.319
89
non-implementation by Member States to one of the strongest levers the EU has financial
incentives.117
3.1.2. Does it require new legislation/policy?

No. This policy option does not entail the creation of a new instrument but rather focuses on
making systematic use of an existing tool. Because it has only recently been developed, there are
no official statistics on how frequently the new powers have been used.
3.1.3. What are the possible challenges or limitations to this option?

The number of legislative measures that could serve as the basis for infringement proceedings
remains limited.118 To illustrate, Anagnostou and Psychogiopoulou (2014) identified only two
EU hard law instruments against corruption the 1995 PIF Convention and the 1997 Corruption
Convention, along with their respective protocols. The point about the limited number of
applicable binding measures is also made by Dori (2015), who stressed that the test of a Member
State failing to implement EU law is not easily fulfilled by non-binding recommendations.
However, some legislation is in place, such as the new public procurement directives, which are
now being transposed by Member States.
In this context, it is also important to recall the role of the Commission as the guardian of the
EU Treaties, as laid out in section 1-I-5.
In an effort to increase its options in the area, short of triggering the Article 7 procedure, the
Commission introduced in 2014 a new framework, intended to address existing threats to rule
of law that cannot be addressed through Member State action or through infringement
proceedings.119 Recently, this process was launched in connection with recent political
developments in Poland (EC 2016). Given the contribution of the fight against corruption to
overall rule of law, this framework may be considered as an addition to the tools at the EUs
disposal.120
When contrasting EUs enforcement powers with those of other bodies, one interviewee pointed
out that the (quasi)voluntary character of mechanisms such as GRECOs evaluations may be
instrumental in receiving buy-in from Member States, who may be reluctant to participate in
projects of a more mandatory nature. This represents a potential limitation on the use of
coercive measures by the EU. It is a dimension to consider when assessing the balance of hard
and soft measures that could be put forth by the EU and other actors.
For instance, there are very specific guidelines in place to determine the applicable financial penalties.
See (EC 2015b).
118 A similar limitation can be found with respect to efforts to ensure an effective implementation of the
Charter of Fundamental Rights in accordance with the Charters Article 47, guaranteeing a remedy in case
of a violation of rights and freedoms. However, the Charters Article 51 explicitly states that this applies to
Member States only when implementing Union law. (FRA, n.d.-a).
119 The process envisaged under this framework consists of three stages: an assessment of the situation by
the Commission, followed by the Commissions recommendations and by monitoring of how the Member
State in question acts on the recommendations. Depending on the outcomes, the situation is either
resolved satisfactorily or Article 7 is invoked (Kochenov & Pech 2015).
120 To illustrate, Butler (2012) discusses the possibility of using EU rule of law competencies to follow up
on indicators from the Justice Scoreboard.
117
PE 579.319
90
3.2. Support new legislation to harmonise protection for

whistleblowing within Member States and/ or provide protections
to whistleblowers within European institutions
-
There is evidence of variability in national-level prevention, control and regulation

mechanisms within Member States.
The effectiveness of the fight against corruption might be reduced by gaps in EU Law (in
particular, the absence of an EU instrument on the protection of whistleblowers).

One course of action is to adopt an EU instrument on whistleblowing with the aim to (i) address
the insufficient protection to whistleblowers afforded by some Member States, or (ii) insufficient
protection to whistleblowers from EU intuitions.
As mentioned above, the Restarting the Future campaign has called for the adoption of an EU
directive on whistleblowing and the establishment of a European Authority for Whistleblowing.
The following possible features of whistleblowing legislation, related in particular to addressing
corruption in regional and local public procurement, were identified in the literature review
and interviews:
-
Develop sectorial legislation: One of the Anticorruption reports recommendations

touched on the point of prevention and detection and called for efforts to raise
awareness about the need and know-how for prevention and detection of corrupt
practices at all levels of public procurement. In this context, an option for EU action
with respect to whistleblowing, as presented by an EP legal service representative, is to
use sectorial legislation, along the lines of offering preferential treatment to contractors
who cooperate with the Commission in cartels cases. 121
Require companies providing services to EU institutions to demonstrate

whistleblowing protections: On a related note, an interviewee suggested that the EU
should require companies who benefit from EU funds to demonstrate that they have
effective whistleblowing policies in place. This would be somewhat akin to provisions
included in the US Sarbanes Oxley Act and subsequent regulations (see Box 7) and
could be considered for inclusion as a legislative measure. In this context, we recall the
adoption of EU Directive 2014/95/EU on the disclosure of non-financial and diversity
information (EU 2014f). The directive will require large companies122 to report
This is frequently refered to as the Commissions leniency policy (EC, n.d.-c).

The criteria include more than 500 employees and a balance sheet in excess of 20m. (Thomas &
Maguire 2014).
121
122
PE 579.319
91
information on a wide range of environmental, social and governance issues, including

any anti-bribery and anti-corruption procedures the companies have put in place.
However, Linklaters (2014) pointed out that the directive leaves a large amount of
discretion to companies in terms of what they decide to report and does not include any
sanctions.123
Box 7: Contribution of the Sarbanes Oxley Act to the
protection of whistleblowers in the United States124
The Sarbanes Oxley Act was passed in 2002, following the scandals of Enron and WorldCom, as
an effort to reform corporate governance and improve financial regulation in the United
States. The majority of its provisions applied only to publicly traded companies but its
measures to protect whistleblowers covered public and private companies alike (Gates 2011).
The act expanded the protection of private-sector whistleblowers, who until then were largely
protected only if reporting concerns related to public health and safety, and broadened the
scope of protected disclosures. The act also introduced stronger penalties and corrective
measures (both civil and criminal) for any reprisals against whistleblowers and made the
burden of proof favourable to employees.
In addition, the adoption of Sarbanes Oxley spurred the enactment of whistleblower protection
at the state level, for instance in California and in Connecticut. With respect to publicly traded
companies, the act requires them to disclose whether they have codes of ethics in place
covering senior executives. Subsequently, the NYSE started requiring companies listed on the
exchange to demonstrate they have ethical codes in place covering all employees, including
rules against retaliation against employees (Westman 2005). Similar requirements for federal
contractors were enacted in the 2007 Federal Acquisition Regulations, which required
contractors with contracts exceeding $5m to have in place a code of ethics and business
conduct, including an employee awareness programme, an internal controls system and an
employee reporting process (Jackson Lewis Corporate Governance Practice Group 2009).
3.2.2. Does it require new legislation/ policy?

Yes. There is currently no EU instrument on whistleblowing
Although Member States may elect to introduce stricter requirements in their transposition of the
directive.
124 The purpose of this box is to provide additional detail with respect to a whistle-blower protection
mechanism suggested by interviewees as an example of good practice. We note that protection of
whistleblowers in the United States has a longer tradition, expressed in, among others, qui tam provisions
in the 1856 False Claims Act, and more recently the Whistle-blower Protection Act of 1989. For a
discussion of these measures see, for instance (Doyle 2009; Shimabukuro & Whitaker 2012).
123
PE 579.319
92

In relation to the Member State level, this option is hampered by the lack of a clear legal basis
for such an action. A representative of the EPs legal service has stressed that there is no legal
basis for a general legislation addressing areas outside of EU activities (EP 2016). Recognising
this, Box 8 sets out a non-legislative alternative that would be within the competency of the EU.
However, the Union has the competence to act with respect to whistleblowing within EU
institutions and for people and entities dealing with EU institutions.
Box 8: Possible non-legislative action to improve
protection for whistleblowers at Member State level: a whistleblowing pact
As an alternative to (or as well as) bringing forward legislation to protect whistleblowers, the
EU could support the development of a whistleblowing pact among Member States, analogous
to successful inter-Member State agreements in other policy areas, such as the Fiscal Compact
(EC, n.d.-k).125 A recommendation for this kind of action was made to the CoE by Stephenson
and Levi (2012), who argued for the formulation of a recommendation based on existing
principles. Advantages of this, they pointed out, would be that it would avoid the laboriousness
of negotiating a stronger legislative instrument and the desirability of letting each jurisdiction
take into account their specific contextual factors.
3.3. Support new legislation to define a public official
-
There is evidence of variability in prevention, control and regulation mechanisms within

Member States.
The effectiveness of the fight against corruption might be reduced by gaps in EU law (in
particular, the absence of common definition of a public official).

Developing legislation to provide a single definition of a public official that would apply across
all Member State and within EU institutions.

Yes. There is currently no legislation providing a common definition of a public official,
The Fiscal Compact is an intergovernmental treaty setting standards and objectives for fiscal and
monetary policy. It forms a part of the broader Treaty on Stability, Coordination and Governance (TSCG)..
125
PE 579.319
93
However, this issue may be resolved pending adoption of a Commission proposal for a
directive on criminal law protection from fraud and related offences to the EU financial
interests (EC 2012c).126
Submitted in 2012, the proposal offers a definition in Article Four of the proposed directive,
stating that a public official includes persons holding a legislative, administrative or judicial
office or otherwise exercising a public service function for the Union or in Member States, as
well as persons exercising such a function in a third country (EC 2012c). This proposed
definition was supported by the follow-up documents from the Council and the Parliament in
2013 and 2015 with the aim to adequately protect Union funds from corruption and
misappropriation and includes everyone assigned a public service in relation to Union funds
(Council of the EU 2013).
However, the proposed directive has not been adopted yet (EUR-lex, n.d.). The fact that the
fraud directives definition would be transferrable to other corruption-related instruments is
exemplified in the latest draft of the EPPO Regulation, which notes that its definitions will need
to be brought in line with those included in the final text of the PIF-Directive (Council of the EU
2015b).

The enforcement of legislation incorporating a unified definition of a public official would still
be heavily dependent on the effectiveness of its implementation by Member States.
3.4. Make improvements to the Anti-corruption report
-
Corruption at the EU level does not fall within the scope of the ACR
There is no formal assessment procedure for the ACR
The ACR includes little new data.

The following possible improvements were suggested to the ACR by interviewees:
Include the EU within the scope of the ACR (this was originally the plan for the ACR).
Replace the ACR with a broader rule of law monitoring framework. Pech (2015) notes that the
ACR exists in parallel with other monitoring tools, such as the Justice Scoreboard, which
Follow-up documents from the Council and the Parliament are general approach of 3 June 2013,
Council Doc. 10232/13 and opinion of the Committee of Legal Affairs, A7-0000/2013, respectively).
126
PE 579.319
94
address related but distinct areas. He suggested it may be more effective to replace this range of
(in his view) weak (Pech 2015) tools with a broader rule of law monitoring framework. This
suggestion is also suggested by, among others, Mller (2015), who discusses the establishment
of a Copenhagen Commission, which would assess Member State progress in a fashion that is
similar to what candidate countries undergo (i.e. progress towards the Copenhagen Criteria).
Involve Member States earlier in the process of developing the ACR: In the 2014 iteration of
the report, one interviewee felt that Member State may have been involved too late in the
process of its compilation. Going forward, involving countries sooner would be beneficial and
the interviewee thought this was being planned for the next (2016) iteration. Greater
involvement with other actors such as NGOs and private sector representatives would also be
helpful.
Raise the profile of the ACR: One interviewee called for the ACR to be used more
courageously as a vehicle for the generation of political will for policy reforms in Member
State. This effort may include working to raise the profile of the ACR and communicating its
findings to wider audiences who could put pressure on individual Member State. How this
could be done remains an open question.
Increase the number of outputs from the ACR: One interviewee suggested that the ACR could
generate more outputs. It may be useful to make information available between biennial
publications of the report, to share all the monitoring data collected and to enable ongoing
review, discussion and sharing of experience. One possible platform for this may be the
experience sharing workshop programme discussed above.
Add more data into the ACR: An interviewee believed that additional data could be brought to
bear in the preparation of the ACR. One possibility would be justice system statistics, ideally
comparable across countries (this is currently being worked on by the European Commission),
though these are inevitably quite time-lagged, because of delays between offending, detection,
investigation and trial, even where all those procedures take place.

This does not require new law or policy.

-
Interviewees noted that the inclusion of the EU in the ACR would not result in an
independent, external review and may thus not be as desirable as other options.
Increasing the scope/ data / outputs would require additional resources and time.
Replacing the ACR with a broader monitoring framework might not be supported by
the Member State.
PE 579.319
95
3.5. Extend aspects of the Cooperation and Verification Mechanism

to other Member States
-
In some instances EU and other legislation has not been formally transposed by
Member States.
There is evidence of variability in prevention, control and regulation mechanisms
within Member States.
The effectiveness of EU action is dependent upon skills, capabilities and capacities
within Member States

Several interviewees agreed that some other countries would benefit from some of the elements
of the CVM. This would involve implementing a monitoring programme for selected Member
States for example, those that fare poorly in the ACR reports or GRECO assessments.

The legal basis for the current CVM can be found in Bulgaria and Romanias accession treaties.
Any application of a CVM-type monitoring mechanism in a non-accession context would
require a new, tailored CVM to be designed, and would require buy-in from the Member State
to which it applies.127

Several interviewees noted the difficulty in making this politically feasible.
A study conducted by the Centre for European Policy Studies (Alegre, Ivanova & Denis-Smith
2009) argued that the same level of EU interference in domestic political affairs may not be
acceptable in other, particularly older Member States. In fact, such efforts may even be
counterproductive and undermine public confidence in the EU area of freedom, security and
justice by confirming suspicions that the tentacles of Brussels are reaching right into the heart
of national sovereignty (Alegre et al. 2009 p.5).
The CVM responded to very concrete needs in Bulgaria and Romania (as reflected in the
different design of the CVM in BG and RO) and any application of CVM elements elsewhere
would need to follow local needs.
As one possibility, Alegre et al. (2009) suggest that a CVM-type model might be applied in situations
where a Member State is found in breach of its rule of law obligations.
127
PE 579.319
96
3.6. Make improvements to the EU Justice Scoreboard
-
The Scoreboard has a limited scope.

-
Increase the scope of the Scoreboard to include criminal courts.
In addition, one interviewee mentioned several other options for improvement, which
are currently being explored by the Commission. For instance, the EC has started
conducting pilot studies to assess the state of play in very specific areas and topics, such
as EU competition law, which may evolve into full-size additions to the scoreboard.
There is also scope for collecting more data and increasing the number of contributing
indicators in the scoreboard, exemplified by a recent effort to gather information on
how well Member States publish court judgments online.

No, amendments to the Scoreboard do not require new legislative or policy instruments.

Increasing the scope to include criminal courts is currently not considered possible due to
insufficient comparable data. Improvements in data collection are a prerequisite for this
expansion.
The other options are currently under consideration by the Commission.
3.7. Take steps for the EU to accede to GRECO
This measure could address the following challenges set out in Section 3II-2:
-
Corruption at the EU level does not fall within the scope of the ACR.
Lack of transparency in relation to law making and lobbying within EU institutions.
Room for improvement in the rules on conflict of interest by EU officials.
PE 579.319
97

One possibility to improve the monitoring of and reporting on corruption at the EU level would
be for the EU to join GRECO and thus become subject to its evaluation procedure, as envisaged
by Article 5 of GRECOs statute (CoE 2015b).
The European Union is collaborating closely with GRECO and has considered acceding to the
group, in parallel to the existing memberships of EU Member States and other European
countries (EC 2012a). A call for the EUs accession to GRECO was also included in the EUs first
Anticorruption Report.
The option of EU accession to GRECO was also highlighted independently by several
interviewees, who felt the EU would strongly benefit from an external review conducted
according to GRECO standards and in line with its principle of equal treatment of members. In
addition, interviewees felt that, with the EU becoming more closely embedded in the
discussions and work of GRECO, there may be positive spill-over effects further downstream at
the level of individual Member State in the form of better coordination of anti-corruption
policies and monitoring of the state of play. GRECO shares this perspective as well, as
evidenced by an explanatory memorandum on the implementation of the Memorandum of
Understanding between the Council of Europe and the European Union. The document
reiterates GRECOs position that EU participation in GRECO would contribute to more coordinated anti-corruption policies in Europe and strengthen their impact (CoE 2014a).128

This option would require the adoption of an agreement on the EUs accession to GRECO.

According to EP representatives, the ability of the EU to accede to GRECO and the conditions
under which this would happen may have influenced Opinion 2/13 of CJEU (CJEU 2014b), that
the EU could not accede to the ECHR as envisaged by the current draft agreement (DouglasScott 2014). In its reasoning, the Court felt that the draft agreement treated the EU as a state,
thus not paying attention to the intrinsic nature of the EU, its law and its autonomy.129
However, the consequence of this ruling with respect to the EUs accession to GRECO remains
unclear. The EUs accession to GRECO was declared a political priority by the EU (Council of
the EU 2015a) and CoE (CoE 2015a) representatives alike even in the aftermath of the 2/13
ruling, although two interviewees noted that a legal analysis of how to proceed was needed.
One interviewee suggested that this appears to be a discussion taking place primarily within
Article 37.
Specifically, the court found that the following fundamental issues were not addressed sufficiently in
the draft agreement: 1) characteristics and autonomy of EU law, 2) article 334 of the TFEU, prescribing MS
to settle disputes pertaining to the application of EU Treaties to means prescribed in those treaties (as
opposed to the ECHR), 3) co-respondent mechanism, envisaged to address how responsibility between the
EU and MS is split, 4) procedure for the prior involvement of the Court, and 5) judicial review in the area
of common foreign and security policy. For a CJEU press release summarising these arguments (see, CJEU
2014a).
128
129
PE 579.319
98
EU institutions, as exemplified by two questions to the Commission on the topic raised by

MEPs in the second half of 2015 (EP 2015f; EP 2015g).130 From the perspective of the Council of
Europe, the EU continues to be invited to accede as soon as possible. (CoE 2015c).131
3.8. Establish a European Public Prosecutors Office
This measure could address the following challenges set out in Section 3II-1:
-
OLAF and other institutions rely on Member States to initiate prosecutions regarding the
use of EU funds.

The Commission has put forward a proposal to set up the European Public Prosecutors Office
(EC 2013e). If approved, this would mean the establishment of an independent office with the
authority to prosecute corruption cases132 independently of Member State prosecution
authorities, albeit in Member State courts.
According to the original legislative proposal, the EPPO was to be a decentralised body of the
Union with its own legal personality.133 As one interviewee explained, by being staffed with
people elected by EU institutions and accountable to the EU (i.e. independent of individual
Member States), the room for national influence on the ultimate decision whether to take a case
forward should be minimised. In addition, the planned modus operandi for the EPPO would
avoid duplication (and the ensuing risk of time barring of cases) of work between OLAF and
Member States as the EPPO would work directly on the basis of Member State information. As
an added benefit, this would enable OLAF to focus on purely administrative cases and leave
criminal investigations to the EPPO.
Over the course of the ongoing negotiation process, the envisaged structure of the EPPO has
been modified. According to the latest available consolidated proposal (Council of the EU
2015b) the EPPO, while retaining its decentralised structure, would be organised at both the
central level and at the decentralised level:
-
At the central level, general oversight responsibilities would be vested in a college

consisting of the European Chief Prosecutor and one European Prosecutor for each
Member State. In addition to the college, the central level would include a permanent
The first question was submitted by Dennis de Jong (GUE/NGL) , Elly Schlein (S&D) , Benedek Jvor
(Verts/ALE), Ignazio Corrao (EFDD) , Ana Gomes (S&D), Marian Harkin (ALDE), and Monica Macovei
(PPE) on 29 September 2015 and the second by Esteban Gonzales Pons (PPE) on 17 December 2015.
131 Article 7.4.
132 As specified in the proposal, this includes lodging the indictment and any appeals until the case has
been finally disposed of. Article 4, (EC 2013e).
133 Article 3, (EC 2013e).
130
PE 579.319
99
chambers made up of a smaller number of college members and tasked with

monitoring and directing investigations and prosecutions.
-
At the decentralised level, the EPPO would consist of Delegated European Prosecutors
who would be located in individual Member States and act there on behalf of the EPPO.
This is very much in line with the original proposal for the EPPO, although the final
scope of the responsibilities of the Delegated Prosecutors is dependent on the final
agreement on the powers of the central level bodies.
According to interviewees, the establishment of the EPPO could result in a range of

improvements in the fight against corruption.134 This conclusion is echoed in an impact
assessment accompanying the EPPO proposal, which concluded that an independent
prosecutors office would constitute an adequate response to the issue of safeguarding the
Unions financial interests (EC 2013a). This could address one of the limitations of the current
operation of OLAF, i.e. the low rate of convictions resulting from national prosecutions.135
One interviewee also suggested that the establishment of the EPPO would make a welcome
contribution to fighting corruption at the local and regional level, particularly in the domain of
public procurement. This is because it is envisaged that the EPPO would have, at least in some
cases,136 a decentralised presence in Member States. Through this organisational arrangement,
the EPPO would have members of staff who are familiar with the national justice systems
embedded at the local level.
Another area where it is hoped that an EPPO would have a positive impact is cross border
crime and fraud in particular.137 This is of particular importance because, as numerous
interviewees pointed out (and as mentioned in section 3-3), currently Member States,
particularly smaller ones, often do not have the capacity and expertise to take on these complex
cases. Furthermore, there remains a problem that it is not clear under the current system what
reason a Member State would have for taking on difficult and expensive cases that might result
in them having to repay fraudulently spent funds to the EC. This is particularly the case if
Member States are not able to recover those funds from the fraudsters, either because this is not
possible through judicial procedures and/or because they have been spent. This issue might be
addressed by an EPPO, but will depend on the final configuration of the Office and, as several
interviewees stressed, on the extent of its independence from Member States this is discussed
below.
In summary, at its maximum, the EPPO will be the first Union body to exercise powers that are
traditionally attributed to nation-states. Its attributions will include conducting criminal
investigations and prosecuting suspects, who are suspected of EU fraud and corruption, on the
territory of participating Member States, against their own citizens, and without Member State
consent. This in itself is a milestone in European integration as it inserts an EU body into the
See also (Franssen, 2013).
We note that the following discussion is speculative since the actual arrangements of the EPPO are still
being considered. Information presented in this section is based primarily on expert interviews, though it
is worth noting that the establishment of an independent EPPO is also one of the recommendations in TIs
Integrity System report.
136 This is particularly applicable to large Member States, though not all of them. For instance, France has
already communicated that they would prefer to have a centralised EPPO presence.
137 As one interviewee noted, from the perspective of the EU, a particularly important issue in this aspect is
carousel crime and resulting VAT evasion, due to its prominent contribution to the EU budget.
134
135
PE 579.319
100
direct relationship between citizens and nation-states, a relationship in which the latter have so
far had exclusive legitimacy of repression (use of violence and delivering justice) in case the
social contract is breached. The EPPO will take over the exercise of some core state functions
(investigate, prosecute and bring cases to court) and thus transform the way in which national
justice systems work in the EU. Criminal prosecutions will gradually become European and the
EPPO's very existence will trigger legal, including constitutional and regulatory, and/or
institutional reforms. The impact is also symbolic: the Union protecting its budget may improve
its standing in public opinion.

Yes, but proposals have already been put forward by the Commission

There are concerns that the proposed collegiate structure dilutes the independence of the
EPPO from Member States and could limit its effectiveness. One interviewee felt the move
towards a collegiate arrangement would be to the detriment of the institutions effectiveness.
The interviewee highlighted Eurojust as an example of the kinds of barriers to effectiveness
faced by a collegiate organisation. This concern about the effectiveness of a collegiate-type
arrangement was also registered in the EPPOs impact assessment, although the actual details
of the college option discussed in the document differ somewhat from the latest proposal text.
Similarly, an EP interim report on the proposal found it regrettable that the collegiate structure
was under consideration.138
Overall, the departure from the original provisions bring in an increased role for Member
States, which raises questions about the extent of EPPOs independence from Member States
(this issue was raised in the EPs interim report (EP 2015b) and at an April 2015 conference on
the EPPO at the European Academy of Law) (Council of the EU 2015c). Similar concerns are
expressed in objections raised by some Member States to the proposed division of
responsibilities and in suggestions to limit the competencies of the Permanent Chambers staffed
by Member State-nominated European Prosecutors (i.e. College members) (Council of the EU
2015b). As an alternative, a number of Member States have suggested that a system where the
European Delegated Prosecutors are responsible for taking the bulk of the operative decisions
would contribute significantly to the effectiveness of the EPPO (Council of the EU 2015b).
Uncertainties also persist surrounding the future relationship between the EPPO and OLAF.
The original proposal afforded the EPPO exclusive jurisdiction in cases related to the protection
of the financial interests of the EU, leaving OLAF with responsibility in administrative
investigations not covered by the EPPOs competencies (EC 2013b).139 However, an EP
resolution from 2015 called for a clarification of the relationship between EPPO and OLAF and
their responsibilities in protecting the Unions interests (EP 2015b).
At the same time, however, the EP agreed the Chambers should play a leading role in EPPOs
activities (EP 2015b). Rizzo (2015) also felt that the collegial structure may be too heavy to be effective.
139 This position was reiterated in a reply to an MEP question in May 2014 (EP 2014a). According to the
original proposal, areas not covered by EPPO competencies within OLAFs remit include crimes
committed by EU staff without a financial impact and irregularities affecting EUs financial interests (EC
2013f).
138
PE 579.319
101
Not all Member States will be covered by the EPPO. Denmark (through its JHA opt-out) is not
involved in the setting up of the EPPO. The UK and Ireland have not opted into the EPPO (EP
2015h). To account for the possibility of a lack of a unanimous Council decision, 140 Article 86
TFEU includes a provision for enhanced cooperation in the event of nine interested Member
States, which would be sufficient for the establishment for the EPPO.
The non-participation of some Member State could also affect the relationship between OLAF
and the EPPO. In this scenario, the EPPO impact assessment foresaw the creation of two groups
of Member States, one where OLAF would continue to be responsible for external
investigations and one where it would not (EC 2013a).
The EPPO would rely on information from other agencies. If implemented as currently
planned, the EPPO would be heavily dependent on information provided by other agencies,
notably justice system agencies in individual Member States. For this reason, current proposals
include providing the EPPO with tools to enforce cooperation by Member State:
-
The ability to recentralise cases at the level of the EPPO and proceed independently of
the Member State authorities.141
The ability of the Chief Prosecutor of the EPPO to talk directly to national police
representatives to compel cooperation.
The ability of the EPPO to take Member States to the CJEU (similarly to the current
powers of the Commission).142
According to one interviewee, there is a strong possibility that additional countries will not subscribe to
the EPPO. As one of the contributing factors, the interviewee pointed out that corruption is still regarded
as a sensitive issue by Member States, which want to keep it away from public domain.
141 According to one interviewee, Member States with a federal system of government are an inspiration
for this arrangement. In such countries it is possible for federal authorities to step in to take a case forward
where local/regional authorities do not intend to do so. This is discussed in Article 22 of the latest
consolidated version of the proposal (Council of the EU 2015b).
142 We note these provisions, as reported by an interviewee familiar with the negotiation process, are a
product of ongoing discussions and are not necessarily reflected in the original proposal for a Council
Regulation establishing the EPPO.
140
PE 579.319
102
CHAPTER 4 COSTS OF NON-EUROPE: WHAT ARE THE POTENTIAL

BENEFITS OF ACTION AT EU-LEVEL?
Chapter summary and key findings

Research activities: This chapter explores the costs of non-Europe in the area of corruption. It
draws on:
The total costs in GDP terms of corruption across the EU-28 Member States, calculated in
Chapter 2.
Evidence on the gaps and barriers in existing measures, set out in Chapter 3.
The potential actions at EU level that might address challenges with current measures, also
set out in Chapter 3.
Methodologies: This chapter employs a quantitative exploratory approach to calculate
potential gains from some of the policy options identified in Chapter 3. The aim is to estimate
what fraction of the overall costs of corruption could be recovered with additional action at the
EU level. It is important to note that calculations in this chapter are based on assumptions,
which we clearly outline where relevant. The analysis provided in this chapter should be
interpreted cautiously. Our approach makes best use of the very limited data available, and
this means that some of the hypothetical situations on which we base the analysis might not
happen in the short or even longer term or might not be politically feasible. Nevertheless, the
calculations in this chapter give a flavour of what could be gained by EU action.
Of the eight possible actions set out in Chapter 3, Section III, we are able to undertake a
quantitative analysis for the following two, based on available data:
Applying aspects of the CVM (or coordinated similar dedicated monitoring mechanism) to
other Member States.
Establishing an EPPO.
In addition, we are able to undertake an analysis of the potential effects of the implementation
of the EU Directives on Public Procurement (especially e-procurement) (described in Chapter 3,
Section 1).
Key findings:
We predict that a CVM-like mechanism applied to more Member States, in addition to
Bulgaria and Romania, could reduce the costs of corruption in GDP terms by around 70bn
annually (or around 8 per cent of the overall costs of corruption)
The establishment of an EPPO could reduce the costs of corruption related to EU funds by
around 200m annually.
The implementation of a full EU-wide e-procurement system could reduce the costs of
corruption risk in public procurement by around 920m each year.
PE 579.319
103
1. Cost of Non-Europe: establishment of a Cooperation and

Verification Mechanism for other Member States
The motivation of this section is to investigate what would happen if Member States were
subject to a mechanism similar in scope and objectives to the CVM. To do this we exploit a
quasi-experiment provided by the fact that the A10 countries (i.e. countries that joined the EU
in 2004) were not subjected to the CVM, whereas Bulgaria and Romania have been since their
2007 accession. In this setting, we investigate how the level of corruption evolved for Bulgaria
and Romania before and after accession and CVM, compared to the Member States in the 2004
enlargement using a differences-in-differences estimation approach.143
As noted elsewhere in this report, we recognise the uniqueness of the CVM as a monitoring tool
and the limitations (both political and practical) surrounding its applicability (or the
applicability of its parts) to other Member States. However, we consider the idea of
implementing a mechanism similar to the CVM to be a useful approximation of the potential
extent of costs stemming from gaps in monitoring mechanisms in general.
1.1.
Methodology: CVM as quasi-experiment
The idea is to compare whether (perceived) corruption in Bulgaria and Romania is significantly
lower after 2007 compared to the pre-accession period. Obviously, in a simple before and after
comparison, any change in the levels of corruption could be driven by changes in economic or
political factors in the post-accession period, rather than by the existence of the CVM. Therefore,
to identify the CVM effect, we compare the change for Bulgaria and Romania with the before
and after change in levels of corruption for the A10 countries, which were not subjected to the
CVM.
It is important to stress that the CVM coincides with EU accession for both Bulgaria and
Romania and therefore it is not straightforward to disentangle the CVM effect from reductions
in corruption levels that might result from other changes in the economic situation and
transition in line with EU Membership. However, we believe that by benchmarking it to the
A10 countries and their post-accession situation we seek to filter out the effect of the CVM on
the levels of corruption. The use of the new Member State joining in 2004 further recognises the
fact that candidate countries were also subject to regular progress reporting by the EU prior to
their accession. This way, we aim to account for any effect EU pre-accession monitoring may
have had in Bulgaria and Romania. In order to take into account different initial corruption
levels in individual countries, we calculate the percentage change of the corresponding
corruption indicator. In essence, we estimate the following equation using the same data as in
chapter 2:
where
corresponds to the (log) level of corruption in country i in year t.
value 1 for countries under CVM (Bulgaria and Romania);

accession period and 0 in the pre-accession period;
143
takes the value 1 in the postis a set of other control variables,
See (Angrist & Pischke 2008) for a summary about the methodology.
PE 579.319
104
takes the
including growth of GDP per capita, trade openness, levels of inequality, the rule of law,
population growth and the level of democratisation (polity2 variable). Since our post-accession
period varies slightly across countries under CVM and those who are not, we include a full set
of year indicator variables. The differences-in-differences estimator the OLS estimate of
, the
coefficient on the interaction between

and
. In essence, if
then the
countries under the CVM lowered their level of corruption more than the A10 countries without
the CVM after accession to the EU.
As already mentioned in Chapter 2, only the ICRG corruption index can be consistently
compared over time, whereas the CPI and the COC are not fully comparable over time. We will
have to consider this when interpreting the results.
The results of estimating equation (3) are highlighted in Table 18 which reports the findings for
all three corruption indices.
Table 18: Effect of CVM on levels of corruption
(1)
(2)
OLS
estimation method:
dependent variable:
post_t
CMV_i
CMV_i*post_t
(3)
ICRG index
CPI index
COC index
0.2133
-0.0381
-0.0103
(0.047)***
(0.034)
(0.035)
0.1612
0.0989
0.1669
(0.065)**
(0.052)*
(0.037)***
-0.1515
0.0234
-0.0072
(0.075)*
(0.084)
(0.077)
Observations
128
Notes: Robust standard errors in parentheses;*** p<0.01, ** p<0.05, * p<0.10. Estimated on a sample of A10 countries
plus Bulgaria and Romania over the years 1996-2014. Each estimation in columns (1)-(3) include the following control
variables: growth GDP per capita, population growth, trade openness, rule of law, democratisation (polity 2) and a full
set of year dummies.
Using this empirical approach, we find for instance that the ICRG corruption index on average
decreased by around 15 per cent more post-accession compared to the A10 countries not subject
to the CVM (column 1). In contrast, we do not find any statistically significant effect of the CVM
on the two other corruption indices (column 2 and 3).
However, it is important to stress again that previous literature strongly suggests using the
ICRG index for any cross-country analysis over time as it is the only index that has been
consistently measured over time (see e.g. Aidt 2011). The methods and measurement for CPI
and COC have changed over time and may not be fully comparable over time. What is more, it
is important to note that the findings do not suggest that the CVM reduces corruption by 15 per
cent, compared to the A10 Member States, but rather that it reduces the level of the ICRG
corruption index, which measures perceived corruption.
In what follows we employ the differences-in-differences estimate for the ICRG index of
column (1) in Table 18 to calculate what amount of lost GDP could be recovered by having a
PE 579.319
105
similar CVM mechanism in place for additional Member States. To that end we apply the
reduction of the ICRG index in the cost calculation framework introduced in Chapter 2.
1.2.
CVM-like mechanism and the Cost of Non-Europe
As highlighted in Table 3 above, the countries with the highest levels of corruption according to
the ICRG corruption index, apart from Bulgaria and Romania, are Latvia, Lithuania, Italy,
Croatia and Greece. To calculate the potential gains from EU action, we calculate how much
lower the overall GDP loss would be if these five Member States reduced their level of
corruption measured by the IRCG index by 15 per cent under a CVM-like mechanism.
Table 19 highlights the potential gains in terms of GDP from putting five countries under a
CVM-like mechanism. Overall, the potential predicted gains of putting Croatia, Greece, Italy,
Latvia and Lithuania under a CVM-like mechanism are in the ballpark of $78bn (70.2bn) or
around eight per cent of the total costs of corruption in GDP terms for the EU-28. It is
important to note that the majority of this relatively large estimate would be driven by Italy.
Looking at the most recent accession country, putting Croatia under such mechanism could
potentially reduce the lost GDP by around $2.2bn (1.98bn) on an annual basis.
Table 19 : Potential gains in GDP terms CVM mechanism
Member State
ICRG index
actual
ICRG index
(under
CVM)
Lost GDP under CVM Scenario 1 (in US

Dollar)
Gain under CVM Scenario 1 (in US Dollar
Croatia
0.728
0.619
6,712,837,966
2,256,822,730
Greece
0.719
0.611
25,862,938,651
8,794,976,996
Italy
0.729
0.620
193,031,992,575
64,784,705,273
Latvia
0.789
0.671
3,282,864,914
1,031,369,644
Lithuania
0.761
0.647
4,563,388,691
1,475,692,705
233,454,022,797
78,343,567,347
Total
Notes: Column ICRG index (under CVM) shows index for the three Member State assuming 15 per cent reduction of the
index. The last column shows the potential gain from the CVM mechanism for each Member State (Lost GDP actual
from Table 6 minus Lost GDP under CVM). We use an exchange rate between the Dollar and Euro of 0.9 to transform
the values into Euros.
It is important to bear in mind that the calculations presented are only predictions based on an
estimated parameter and depend on the hypothetical case that it would be possible to put these
countries under a strict monitoring mechanism. In addition, one should note that the estimated
parameter of the effect of the CVM is measured over an average of seven years after
implementation. Hence, it is important to stress that the improvements amid the CVM may not
last forever and may plateau at some point, with diminishing gains after that.
2. Cost of Non-Europe: EPPO

We identified the introduction of the EPPO as one of the potential options where action at EUlevel could reduce the overall cost of corruption. The existing impact assessment of the EPPO
(EC 2013a) uses as working assumption that a ten per cent increase in the number of convictions
will lead to a one per cent decrease in the annual fraud related damages suffered, which
represents a combined effect of deterrence and actual recuperation of damages. We use this
PE 579.319
106
elasticity to calculate the predicted reduction in the cost of corruption that would accrue if an
EPPO was created.
2.1. Methodology: higher prosecution rates related to EPPO

Building on data on Member State-level indictment and conviction rates, as reported by OLAF
and the EC (EC 2013f), the working assumption is that the number of prosecutions would be
higher under EPPO, and could possibly be equal to the total number of judicial
recommendations made by OLAF.
Subsequently a higher number of prosecutions would lead to more convictions, under the
assumption that national Member State conviction rates, as share of all cases brought to courts,
would stay the same. Table 20 highlights the current cases brought to court and the conviction
rates at Member State level. Columns 2 to 5 report the current state of actions transferred to
Member States and the corresponding judicial decisions and subsequent convictions taken by
each Member State. Columns 6 to 9 report the hypothetical judicial actions and convictions that
potentially would accrue under EPPO. Note, that the calculations are based on the assumption
that EPPO would bring all actions transferred to a judicial decision under the same conviction
rate as before the EPPO. The higher number of judicial decisions therefore solely drives the
increase in conviction rates. We note that there may be reasons to expect future changes in
conviction rates, for example due to possible improved management of prosecutions; however,
in the absence of relevant estimates, we are not in a position to project any change in conviction
rates as a share of total cases brought to court.
In order to calculate the reduced damages of corruption for each Member State and the EU as a
whole, we use the predicted conviction rates under EPPO and the corresponding %-reduction
in damages, based on the assumption that a ten per cent increase in conviction rates reduces
damages by around one per cent. Since the EPPOs purview would be the protection of EUs
financial interests, we use data on all allocated EU Structural and Investment Funds for the
years 2014-2020144 alongside the assumption that around two per cent of these are at risk of
corruption,145 to calculate the overall reduction in damages related to EU funds.
2.2. EPPO and the Costs of Non-Europe

Table 21 highlights the predicted reduction in the costs of corruption after the establishment of
EPPO. Our findings predict that if all Member States established EPPO, around 200m of the
EU budget could be recuperated per year. It is important to stress that this estimate is likely to
be an upper bound estimate as its calculations are based on the total allocated funds per year for
each Member State, but not all of the allocated funds may actually be realised.
These include the European Regional Development Fund (ERDF), the European Social Fund (ESF), the
Cohesion Fund (CF), the European Agricultural Fund for Rural Development (EAFRD), and the European
Maritime
&
Fisheries
Fund
(EMFF).
Data
downloaded
from
https://cohesiondata.ec.europa.eu/dataset/ESIF-FINANCE-DETAILS/e4v6-qrrq [As of 23 February
2016].
145 Based on an estimation by the European Union (EP 2015j).
144
PE 579.319
107
Table 20: Actions transferred to Member States and subsequent judicial decisions and
convictions before EPPO and after EPPO
Member
States
Actions
transferred
to Member
States
Actions
with
judicial
decision
before
EPPO
Convictions
before
EPPO
Austria
Convictions
as % of
actions
with
judicial
decision
without
EPPO
100%
Belgium
56
28
18
Bulgaria
37
14
Cyprus
Czech
Republic
Actions
with
judicial
decision
under
EPPO
Predicted
convictions
under
EPPO
Increase of
convictions
under
EPPO
compared
to before
100.00%
64.30%
56
36
100.04%
42.90%
37
16
164.55%
NA
0.00%
23
25%
23
187.50%
Denmark
33.30%
33.32%
Estonia
100%
200.00%
Finland
12
12
11
91.70%
12
11
0.04%
France
29
12
75%
29
22
141.67%
Germany
168
114
65
57%
168
96
47.32%
Greece
86
26
19.20%
86
17
230.24%
Hungary
10
0%
10
0.00%
Ireland
NA
0.00%
Italy
112
37
14
37.80%
112
42
202.40%
Latvia
NA
0.00%
Lithuania
88.90%
0.01%
Luxembourg
100%
100.00%
Malta
NA
0.00%
Netherlands
29
16
31.30%
29
81.54%
Poland
90
17
35.30%
90
32
429.50%
Portugal
21
66.70%
21
14
133.45%
Romania
225
128
30
23.40%
225
53
75.50%
Slovakia
16
0%
16
0.00%
Slovenia
NA
0.00%
Spain
54
0%
54
0.00%
Sweden
100%
25.00%
United
Kingdom
19
13
23.10%
19
46.30%
Notes: data based on Olaf and the European Commission. Columns 2 to 5 report the current state of actions
transferred to Member States and the corresponding judicial decisions and subsequent convictions taken by each
Member State. Columns 6 to 9 report the hypothetical actions and convictions under EPPO. The calculations are
based on the assumption that EPPO would bring all actions transferred to a judicial decision under the same
conviction rate as before EPPO. The increase in conviction rates is therefore driven by the higher number of
judicial decisions. The calculated %-reduction in damages is based on the assumption that a ten per cent increase
in conviction rates reduces damages by around one per cent.
PE 579.319
108
Table 21: Predicted annual reduction in the costs of corruption after establishment of EPPO
Member State
Allocated EU
Funds per year
At risk of
corruption (2%)
per year
% reduction in
damages EU
Funds with
EPPO per year
Lower cost of
corruption with EPPO
per year
Austria
703,791,043
14,075,821
10.00%
1,407,582
Belgium
387,940,381
7,758,808
10.00%
776,226
Bulgaria
1,411,955,775
28,239,116
16.46%
4,646,746
Cyprus
130,004,633
2,600,093
0.00%
Czech Republic
3,457,811,808
69,156,236
18.75%
12,966,794
Denmark
179,925,789
3,598,516
3.33%
119,903
Estonia
638,983,212
12,779,664
20.00%
2,555,933
Finland
541,898,222
10,837,964
0.00%
394
France
3,826,936,947
76,538,739
14.17%
10,842,988
Germany
3,982,015,001
79,640,300
4.73%
3,768,824
Greece
2,915,586,546
58,311,731
23.02%
13,425,693
Hungary
3,573,460,486
71,469,210
0.00%
Ireland
479,947,561
9,598,951
0.00%
Italy
6,110,154,574
122,203,091
20.24%
24,733,906
Latvia
804,843,638
16,096,873
0.00%
Lithuania
1,199,258,467
23,985,169
0.00%
300
Luxembourg
20,023,994
400,480
10.00%
40,048
Malta
119,040,167
2,380,803
0.00%
Netherlands
246,901,702
4,938,034
8.15%
402,647
Poland
12,288,151,824
245,763,036
42.95%
105,555,224
Portugal
3,685,266,285
73,705,326
13.35%
9,835,976
Romania
4,407,109,242
88,142,185
7.55%
6,654,735
Slovak Republic
2,190,047,277
43,800,946
0.00%
Slovenia
553,894,115
11,077,882
0.00%
Spain
5,343,517,819
106,870,356
0.00%
Sweden
521,730,983
10,434,620
2.50%
260,865
United Kingdom
Total
2,348,865,927
46,977,319
4.63%
2,175,050
200,169,834
PE 579.319
109
3. Costs of Non-Europe: e-procurement

The 2014 ACR identified the potential of the establishment of a fully functioning e-procurement
system as an option to reduce corruption. Accordingly, measures in support of wider
implementation of e-procurement are an integral part of the 2014 suite of public procurement
directives. It is documented that e-procurement can increase transparency and accountability of
the procurement process by enhancing relationships between public officials and citizens
tracking their activities and improving monitoring and control mechanisms to reduce the
potential for corrupt behaviour (Neupane et al. 2012). E-Procurement is seen as tool for
reducing corruption and increasing integrity in public procurement systems. Globally there
have been concrete attempts to implement full e-procurement systems in South Korea and more
recently in Albania (Transparency International 2011). South Korea was one of the first
countries globally to implement a full e-procurement in 1997 and is nowadays seen as one of the
best practice examples in implementing a system that covers all steps of the procurement
process. At that time, the South Korean government sought to reform its complicated, nontransparent and corrupt public procurement system (World Bank 2004). The example of South
Korea shows that such a system can improve efficiency (it is estimated that it created around
$2.5bn a year in savings) as well as enhancing public trust and reducing corruptive behaviour
by reducing the contacts between officials and suppliers (World Bank 2004).
Overall, the empirical evidence on the effectiveness of e-procurement systems is scarce so far.
Some examples where e-government initiatives improved levels of corruption are documented
for a program implemented in Chile, called Chile Compra. This was analysed in the Chilean
chapter of TI (Garcia-Murillo 2013). An evaluation of the program showed that from 2004 to
2006, perceptions of corruption were lower for government purchases done through Chile
Compra. What is more, some evidence from India reveals the effectiveness from different egovernment projects in combatting corruption (Bhatnagar 2009). The results showed that there
was a reduction in bribes paid between ten per cent to 50 per cent, depending on the project.
Some more recent studies show the effect of e-government on the reduction in corruption using
cross-country comparisons. Garcia-Murillo (2013) for instance, finds that government web
presence reduces the perceptions of corruption. Mistry & Jalal (2012) investigate the
relationship between e-government and corruption in developed and developing countries.
Their results suggest that the use of e-government decreases corruption. Specifically, they
estimate that a one per cent increase in the e-government development index (EDI) (UN, n.d.)
reduces corruption by around 1.16 per cent, whereas corruption is measured by the CPI index.
In what follows, we use a similar approach like Mistry & Jalal (2012) but link the EDI index to
the corruption risk index (CRI) introduced in Chapter 2.
3.1. Methodology: e-procurement and corruption

To assess the potential gains of introducing an EU-wide full e-procurement system, we draw on
data from the EDI index. The EDI index is provided by the United Nations Public
Administration Network, which conducts a bi-annual e-government survey including a section
on e-government readiness (UN 2014). The index represents a comparative ranking of countries
according to two primary components: (a) the state of e-government readiness; and (b) the
extent of e-participation. It is important to stress that the index measures the extent and
participation of a countrys e-government and hence not only the e-procurement system. We
therefore have to assume implicitly that a Member States achieved level in the EDI index
PE 579.319
110
corresponds to its (unobserved) e-procurement readiness and participation. Nevertheless, South

Korea, the country that serves as the best practice example in e-procurement is ranked number
one in the EDI index, therefore we assume that there is at least a strong correlation between the
measure of e-government and e-procurement. Given that South Korea is seen as best practice
example in e-procurement, we utilise the EDI index for South Korea as benchmark and calculate
the difference of each Member States index level to the one of South Korea. This difference gives
us an indication or proxy on how far away the current system in a Member States is from the
realisation of a full e-procurement system.
We make use of the publicly available yearly cross-sections of the EDI index for the years 2010,
2012 and 2014 (UN, n.d.) and link this information to the PP database outlined in Chapter 2. We
aggregate information from the PP database on the country level (average CRI index, average
share of contracts that included EU funds and the average number of contract) for the years
2010, 2012 and 2014. In addition we make use of information provided by the QoG database,
which we used already in Chapter 2 to analyse the costs of corruption at the EU-level (GDP per
capita, share of women in parliament, level democratisation measured by polity2 variable, press freedom,
presidentialism and the gini-index as measure of inequality). Using this data we estimate with OLS
the correlation between average level of corruption risk in public procurement in each Member
State and the EDI index using the following equation:
Where
represents the average corruption risk in public procurement in year t146 and
Member State i.
corresponds to the EDI index and
represents a vector of control
variables, including GDP per capita, the average share of contracts including EU funds, the
average number of procurement contracts, the share of women in parliament, presidentialism,
personalism, press freedom (see Chapter 2 for more details) and the level of democracy
(measured by polity2). represent year effects.
The results of estimating equation (4) with OLS are presented in Table 22. Column 1 reports the
association between the EDI index and the CRI index at the Member State level. It shows a
negative relationship meaning that the higher a country is in the EDI index, the lower the levels
of corruption risk in public procurement. Column 2 reports the results from a model including a
number of control variables. The estimated parameter is lower in magnitude but still
statistically significantly and negative. The finding suggests that a one-unit increase of the EDI
index reduces the CRI index (at the Member State level) by 0.29.147
In what follows, we apply this 0.29 estimate to predict how much lower the corruption risk in
public procurement potentially could be in each Member State with the establishment of a full
e-procurement system. We use the distance from the EDI index of South Korea for each
Member State as proxy on how far away each Member States system is from a full eprocurement system.
Table 4-6 reports the level of the corresponding EDI index for each of the 28 Member States and
South Korea, which serves as the best-practice benchmark. Using the assumptions outlined
above, we multiply the distance between a Member States EDI index and the one of South
Korea with the average negative correlation between the EDI and the CRI index of 0.29 and
146
147
We use the three years 2010, 2012 and 2014 for the analysis.
Both indices, EDI and CRI are on the interval between [0,1].
PE 579.319
111
hence derive the predicted reduction in the CRI index with a full e-procurement system. Table
23 highlights that this predicted reduction in CRI corruption levels would be highest in the
Member States Bulgaria, Romania, Cyprus, Czech Republic, Slovakia and Malta. In order to
predict the reduction in losses in EU public procurement we feed back the new CRI level into
the calculations made in Chapter 2.
Table 22: correlation between e-government index (EDI) and
public procurement corruption risk index (CRI)
(1)
estimation method:
dependent variable:
(2)
OLS
OLS
CRI Index (average by Member State and year)
EDI index
Observations
-0.4932
-0.2996
(0.095)***
(0.146)**
81
81
R-squared
0.2084
0.5669
Notes: (Panel-)Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.10. The results are estimated
on a sample of all EU-28 Member States and the years 2010, 2012 and 2014. The model in column 1 includes only
the EDI index as independent variable. The model in column 2 controls for the level of GDP per capita, press
freedom, share of women in parliament, presidentialism, personalism, level of democracy (polity2), the number of
public procurement contracts and the share of contracts where EU funds are included. In addition, the model in
column 2 includes a regional dummy (for EU15 countries) and year fixed effects.
3.2. E-procurement and the Costs of Non-Europe

Table 24 reports the predictions of applying the expected reduction in corruption due to the
establishment of a full EU-wide e-procurement system to the corruption risk index (CRI)
introduced in Chapter 2. We calculate how much we would expect the CRI index to be reduced
in light of such a procurement system.
In essence, our findings predict that the implementation of a full e-procurement system could
reduce the costs of corruption risk in public procurement by around 924m annually which
corresponds to a reduction of almost 20 per cent of the current costs.
Table 23: Predicted reduction in the corruption levels with the introduction of e-procurement
Member State
E-Government
Index 2014
% Diff to Korea as
benchmark
predicted
reduction in CRI
index
Austria
0.7912
0.1550
0.0449
Belgium
0.7564
0.1899
0.0551
Bulgaria
0.5421
0.4041
0.1172
Croatia
0.6282
0.3181
0.0922
Cyprus
0.5958
0.3505
0.1016
Czech Republic
0.6070
0.3393
0.0984
PE 579.319
112
Member State
E-Government
Index 2014
% Diff to Korea as
benchmark
predicted
reduction in CRI
index
Denmark
0.8162
0.1300
0.0377
0.8180
0.1283
0.0372
0.8449
0.1013
0.0294
0.8938
0.0524
0.0152
0.7864
0.1598
0.0464
0.7118
0.2345
0.0680
0.6637
0.2825
0.0819
0.7810
0.1652
0.0479
0.7593
0.1869
0.0542
0.7178
0.2285
0.0663
0.7271
0.2191
0.0636
0.7591
0.1871
0.0543
0.6518
0.2944
0.0854
0.8897
0.0566
0.0164
0.6482
0.2980
0.0864
0.6900
0.2563
0.0743
0.5632
0.3831
0.1111
0.6148
0.3315
0.0961
0.6505
0.2957
0.0858
0.8410
0.1053
0.0305
0.8225
0.1237
0.0359
Estonia
Finland
France
Germany
Greece
Hungary
Ireland
Italy
Latvia
Lithuania
Luxembourg
Malta
Netherlands
Poland
Portugal
Romania
Slovakia
Slovenia
Spain
Sweden
0.0768
United Kingdom
0.8695
0.0223
Best practice benchmark:
Republic of Korea
0.9462
Notes: the assumptions behind table entries is that South Korea serves as best practice example in terms of eprocurement and that we can infer directly from the level of the EDI-index to how far a Member State is away
from having a similar e-government system like South Korea. This implicitly implies the assumption that the
rank in the EDI index reflects the level of the e-procurement system. What is more, the predicted reduction in
corruption is based on the elasticity that one per cent increase in the EDI index reduces corruption by 1.16 per
cent.
Table 24: Predicted reduction in the corruption levels with the introduction of e-procurement
0.194
Cost Total
(EURO) per year
before eprocurement
33,158,751
Cost Total
(EURO) per year
after eprocurement
26,992,734
0.174
0.119
43,309,315
29,698,149
Bulgaria
0.326
0.209
13,839,039
8,856,914
Cyprus
0.58
0.488
17,590,453
14,802,719
Czech Republic
0.338
0.236
399,601,343
279,225,663
Member State
CRI index
before eprocurement
CRI index after

eprocurement
Austria
0.239
Belgium
PE 579.319
113
Germany
0.313
0.215
140,316,050
96,360,442
Denmark
0.182
0.144
25,580,679
20,304,748
Estonia
0.349
0.312
27,078,887
24,217,515
Spain
0.251
0.222
251,734,281
222,060,087
Finland
0.327
0.312
34,727,318
33,129,445
France
0.224
0.178
72,110,970
57,243,004
Greece
0.292
0.224
77,110,065
59,136,271
Croatia
0.584
0.502
74,003,746
63,633,708
Hungary
0.259
0.211
243,329,958
198,193,455
Ireland
0.224
0.170
3,674,459
2,787,297
Italy
0.371
0.305
696,029,492
571,892,665
Lithuania
0.498
0.434
11,799,625
10,282,060
Luxembourg
0.132
0.078
2,897,453
1,703,129
Latvia
0.27
0.185
88,031,718
60,133,463
Malta
0.382
0.366
706,756
692,880
Netherlands
0.302
0.216
40,078,199
28,615,047
Poland
0.433
0.359
1,417,094,143
1,174,077,645
Portugal
0.314
0.203
39,249,233
25,390,556
Romania
0.449
0.353
368,013,805
289,250,831
Sweden
0.147
0.061
19,118,372
7,975,447
Slovenia
0.346
0.315
39,443,543
35,951,434
Slovakia
0.352
0.316
129,980,618
116,731,679
United Kingdom
0.305
0.283
1,025,876,827
952,026,159
5,335,485,098
4,411,365,145
EU-28 Total

The measurement of the potential benefits of more action at EU-level is marked by significant
challenges surrounding the assumptions made in calculating the predicted cost savings of three
different options, a CVM-like mechanism, the establishment of the EPPO and the potential
benefits of fully implemented EU-wide e-procurement system. Wherever possible we seek to
make assumptions that sound plausible in the context of the subject and are potentially
validated by previous research. Nevertheless, the predictions made in this chapter are only as
valid as the assumptions made in order to derive them.
Furthermore, the chapter focuses primarily on the potential benefits of further action at EUlevel in the area of corruption, but in analysing the potential gains implies there is no detailed
consideration of the direct or indirect costs of further action at EU-level. In essence, for each of
the three options we calculate the potential gross gains per year, but there would be costs of the
implementation. We were not able to fully assess the costs of implementation at this stage. This
should be taken into consideration when interpreting the results of this study. When
conducting a full assessment of the costs and benefits of further EU-action and its added value,
direct and indirect costs of implementation and benefits should be taken into account.
PE 579.319
114
CHAPTER 5 REPORT SUMMARY AND CONCLUSIONS

This study sought to quantify the economic, social and political costs of corruption in the
European Union, investigate gaps and barriers in the existing regulatory framework that hinder
the effectiveness of measures to combat corruption in the EU and identify the potential for
action at EU level that might add value and address the challenges identified.
The focus of the study was on anti-corruption measures relevant to the LIBE Committee of the
European Parliament. The research team employed a mix of quantitative (compilation of
bespoke data set and econometric modelling) and qualitative (review of relevant documents
and interviews with 17 stakeholders) methods. The findings of the study are summarised
below.
1. The costs of corruption in the EU

The study examined a variety of existing high-level estimates of the costs of corruption at the
EU-level and globally. These estimates are based on methodologies that are not always clearly
documented, which means that it is difficult to understand the underlying components taken
into account in generating estimates. The added value of this study is the production of a new
estimate of the costs of corruption that is based on sound methodologies (previously applied in
the academic literature) which have been clearly documented (see Chapter 2).
Under three scenarios (which differ in the assumption made about the scale of reductions in the
level of corruption feasible for Member States in the short, medium and long term), our findings
suggest that corruption costs the EU between 179bn to 990bn in GDP terms on an annual
basis (depending on the corruption measure applied). Unlike some of the existing high-level
cost of corruption estimates, these figures include both direct and indirect effects of corruption.
What is more, the findings suggest that corruption has significant social costs (more unequal
societies, higher levels of organised crime and weaker rule of law) and political costs (lower
voter turnout in national parliamentary elections) and lower trust in EU institutions.
When looking in more detail at a specific sub-sector of the EU economy and public
procurement, the findings of the empirical analysis further suggest that the cost of corruption
risks in EU public procurement are around 5bn per year overall, including most sub-sectors of
public procurement and contracts of all EU Member States.
2. What are the gaps and barriers in the existing regulatory

framework that hinder the effectiveness of measures to
combat corruption in the European Union?
While anti-corruption policy is primarily a Member State competency, the EU has some scope to
act in the spheres of the establishment of the area of freedom, security and justice. The fight
against corruption has featured as an important element of EU justice and home affairs policy.
There are a range of anti-corruption measures applicable to EU Member States and EU
institutions.
PE 579.319
115
The EU has adopted a number of legislative provisions in key conventions and directives
(described in Chapter 3, Section I). Additionally, OECD and UN conventions have addressed
the issue of corruption. The effectiveness of existing legislation is limited by several factors. In
some instances, international measures have not been properly transposed by some Member
States. In other cases, existing measures are not of binding character or there are limits to their
enforceability. In instances where Member States have formally acted in accordance with
international norms and recommendations, there is a substantial variability in the extent to
which these efforts are effective. This may be due to differences in the quality of
implementation and enforcement at the Member State level, which in turn may be a reflection of
differences in factors such as political will or administrative capacity. In some cases, such as the
recent procurement directives, it is too early to assess their effectiveness given the recent nature
of their adoption. In addition, some legislative gaps persist, such as the lack of an EU-wide
system of whistleblower protection or the absence of a harmonised definition of a public
official.
Beyond legislation, a number of monitoring mechanisms exist which aim to record the extent to
which Member States laws and institutions are in line with good practice in the fight against
corruption. These mechanisms largely rely on voluntary participation by Member States and are
intended to improve standards by providing guidance and encouraging reform. The three most
high-profile monitoring mechanisms are the EU ACR, the Cooperation and Verification
Mechanism, and the EU Justice Scoreboard. Additionally, the Council of Europe hosts GRECO,
which also includes Member State-level monitoring. These mechanisms are generally
considered to have contributed to the effectiveness of the fight against corruption, although a
number of areas for improvement are identified by this study based on a review of literature
and interviews with expert stakeholders. One notable limitation is that corruption at the EU
level (i.e. within EU institutions) is not subject to monitoring either as part of the EU
Anticorruption Report or GRECO.
3. Is there a potential for action at EU level and can the gains

for these measures be quantified?
Our study identifies the following possible actions that could lead to potential benefits for the
EU as a whole:
1.
Make use of infringement proceedings against Member States who have not implemented
EU law in relation to the fight against corruption.
2.
Support new legislation to harmonise protection for whistleblowing within Member States
and/ or provide protections to whistleblowers within European Institutions.
3.
Support new legislation to define a public official.
4.
Make improvements to the ACR.
5.
Extend aspects of the Cooperation and Verification Mechanism to other Member States.
6.
Make improvements to the EU Justice Scoreboard.
7.
Take steps for the EU to accede to GRECO.
8.
Establish a European Public Prosecutors Office.
PE 579.319
116
As with the limitations of current measures, these possible areas for action were identified
based on interviews and a literature review.
For two of these options (extend aspects of the Cooperation and Verification Mechanism to
other Member States and establish a European Public Prosecutors Office) we were able to
undertake quantification of the possible gains from more action at EU-level. In addition, we
undertake an estimation of the possible gains from a common EU approach to E-procurement,
as envisaged under the 2014 public procurement directives.
While every care was made to source the best available data, the evidence base on corruption is
limited and there are little relevant and good-quality data on which to build robust calculations.
The estimates provided are necessarily based on a series of assumptions about key factors that
will mediate the impact of these measures.
We predict that applying aspects of a CVM-like mechanism to five more Member States (it is
currently applied only to Romania and Bulgaria) could reduce the costs of corruption in GDP
terms by around 70bn annually (or around eight per cent of the overall costs). According to
our calculations, a large part of this reduction would come from applying aspects of the CVM to
Italy. If the measure were only applied to Croatia (the most recently acceding Member State),
the potential gains would be around 2 bn previously lost GDP per annum. Again, we reiterate
here that these estimates are highly dependent on the underlying assumptions, which are
clearly set out in Chapter 4, Section I.
The establishment of EPPO could reduce the costs of corruption related to EU funds by around
0.2bn annually. This is based on a number of assumptions - importantly, that all Member
States would sign up to an EPPO. Accordingly, this estimate should be treated as an upper
bound estimate indicating the gains that could be made in a best case scenario. Other key
assumptions and limitations of this estimate are explained in Chapter 4, Section II.
Lastly, our estimates suggest that the implementation of a full EU-wide e-procurement system
could reduce the costs of corruption risk in public procurement by around 920m each year.
The method underlying this approach is explained in Chapter 4, Section III and uses South
Korea as a best practice example of a country that has an established and fully working eprocurement system.
4. Implications and unanswered questions

This study has investigated the cost of non-Europe in relation to corruption. Compared to other
policy areas, corruption, being an illicit activity, is an extremely challenging field to apply the
cost of non-Europe methodology. The EU has a limited competence to act to harmonise Member
States law in relation to corruption, and even where EU law exists and has been transposed, the
implementation of anti-corruption depends on a complex range of political, legal, cultural and
institutional factors at the Member State level. Additionally, there is very little conclusive
evidence about the effectiveness of law and policy measures in reducing corruption.
Throughout this study we have explained the limitations of our approach and detailed the
methods applied in calculating the costs of corruption and the potential gains from increased
PE 579.319
117
action by the EU. It is hoped that if more data become available and better indicators emerge,
future researchers will be able to build on the estimates produced in this paper to improve
understanding about how corruption can better be tackled at Member State and EU level.
PE 579.319
118
REFERENCES
Acemoglu, Daron, Simon Johnson, and James A Robinson. 2000. The colonial origins of
comparative development: An empirical investigation, National bureau of economic research.
Aghion, Philippe, Nicholas Bloom, Richard Blundell, Rachel Griffith, and Peter Howitt. 2002.
Competition and innovation: An inverted U relationship, National Bureau of Economic
Research.
Aidt, Toke S. 2011. Corruption and sustainable development. International Handbook on the
Economics of Corruption. 1.
Alegre, Susie, Ivanka Ivanova, and Dana Denis-Smith. 2009. Safeguarding the Rule of Law in an
Enlarged EU: The Cases of Bulgaria and Romania: Centre for European Policy Studies.
Alexeeva, Victoria, Cesar Queiroz, and Satoshi Ishihara. 2011. Monitoring road works contracts
and unit costs for enhanced governance in europe and central asia.
Anagnostou, D, and E. Psychogiopoulou. 2014. Background report on international and European
law against corruption, Anticorrp. 10.1. As of 23 February 2016:
http://anticorrp.eu/publications/d10-1-background-report-on-international-and-europeanlaw-against-corruption/
Anderson, Christopher J, and Yuliya V Tverdova. 2003. Corruption, political allegiances, and
attitudes toward government in contemporary democracies. American Journal of Political
Science, Vol. 47, No. 1. 91-109.
Andersson, Staffan, and Paul M Heywood. 2009. The politics of perception: use and abuse of
Transparency International's approach to measuring corruption. Political Studies, Vol. 57.
No. 4. 746-767.
Angrist, Joshua D, and Jrn-Steffen Pischke. 2008. Mostly harmless econometrics: An empiricist's
companion, Princeton, NJ:: Princeton University Press.
Arrow, Kenneth, Partha Dasgupta, Lawrence Goulder, Gretchen Daily, Paul Ehrlich, Geoffrey
Heal, Simon Levin, Karl-Gran Mler, Stephen Schneider, and David Starrett. 2004. Are
we consuming too much?, The Journal of Economic Perspectives, Vol. 18. No. 3. 147-172.
Babu, R Rajesh. 2006. The United Nations convention against corruption: A critical overview.
Available at SSRN 891898
Bck, Hanna, and Axel Hadenius. 2008. Democracy and State Capacity: Exploring a JShaped
Relationship. Governance. Vol. 21 No. 1. 1-24.
Barroso, J.M.D. 2012. State of the Union 2012 Address. As of 23 February 2016:
http://europa.eu/rapid/press-release_SPEECH-12-596_en.htm
PE 579.319
119
Beck, Thorsten, George Clarke, Alberto Groff, Philip Keefer, and Patrick Walsh. 2001. New
tools in comparative political economy: The Database of Political Institutions, The World
Bank Economic Review. Vol. 15. No. 1. 165-176.
Behr, N. 2015. Germany: Draft law to expand provision on commercial bribery and bribery of
foreign public officials. Global Compliance News. 4 February 2015.
Bhatnagar, S. 2009. Does E-government reduce corruption: Findings from Eight Indian
projects.
Bloom, David E, Jeffrey D Sachs, Paul Collier, and Christopher Udry. 1998. Geography,
demography, and economic growth in Africa. Brookings papers on economic activity. Vol.
1998. 207-295.
Buscaglia, Edgardo. 2003. Controlling organized crime and corruption in the public sector.
Forum on crime and society.
Butler, I. 2012. The Justice Scoreboard: An Effective Mechanism to Enforce Respect for the Rule
of Law? Open Society Foundations, 13 December 2012. As of 23 February 2016:
https://www.opensocietyfoundations.org/voices/justice-scoreboard-effective-mechanismenforce-respect-rule-law
Campos, Nauro F, Ralitza D Dimova, and Ahmad Saleh. 2010. Whither corruption? A
quantitative survey of the literature on corruption and growth.
Carrera, S., E. Guild, and N. Hernanz. 2013. Rule of law or rule of thumb? A New Copenhagen
Mechanism for the EU.
Charron, Nicholas. 2016. Do corruption measures have a perception problem? Assessing the
relationship between experiences and perceptions of corruption among citizens and
experts. European Political Science Review, Vol. 8. No. 01. 147-171.
Charron, Nicholas, and Victor Lapuente. 2010. Does democracy produce quality of
government? European Journal of Political Research, Vol. 49. No. 4. 443-470.
Christiane, Arndt. 2006. Development Centre Studies Uses and Abuses of Governance Indicators.
OECD Publishing.
CJEU. 2014. The Court of Justice delivers its opinion on the draft agreement on the accession of
the European Union to the European Convention for the Protection of Human Rights and
Fundamental Freedoms and identifies problems with regard to its compatibility with EU
law. As of 23 February 2016:
http://curia.europa.eu/jcms/upload/docs/application/pdf/2014-12/cp140180en.pdf
, Opinion 2/13 of the Court. 2014b. ECLI:EU:C:2014:2454. As of 23 February 2016:
http://curia.europa.eu/juris/document/document.jsf?docid=160882&doclang=EN
CoE, Civil Law Convention on Corruption, 1999.
PE 579.319
120
, Recommendation Rec(2003)4 of the Committee of Ministers to Member States on

common rules against corruption in the funding of political parties and electoral
campaigns, 2003.
, GRECO Rules of Procedure. Adopted by GRECO at its 1st plenary meeting (Strasbourg,
4-6 October 1999), 2011. As of 23 February 2016:
https://www.coe.int/t/dghl/monitoring/greco/documents/Greco(2011)20_RulesOfProcedur
e_EN.pdf
, The implementation of the Memorandum of Understanding between the Council of
Europe and the European Union. 2014a. As of 23 February 2016:
http://assembly.coe.int/nw/xml/XRef/Xref-XML2HTML-en.asp?fileid=21341&lang=en
, Protection of whistleblowers: Recommendation CM/Rec(2014)7 and explanatory
memorandum.
, 68th GRECO Plenary Meeting (Strasbourg, 15-19 June 2015): Summary Report.
, GRECO. 2014 Group of States against Corruption: 15th General Activity Report. Council of
Europe. 2015b. As of 23 February 2016:
http://www.coe.int/t/dghl/monitoring/greco/documents/2015/Greco(2015)1_GAR2014_EN
.pdf
, The implementation of the Memorandum of Understanding between the Council of
Europe and the European Union. 2015c.
, Etat des signatures et ratifications du trait 173 (Convention pnale sur la corruption)
n.d.-a. As of 23 February 2016:
http://www.coe.int/fr/web/conventions/full-list/-/conventions/treaty/173/signatures
, Etat des signatures et ratifications du trait 174 (Convention civile sur la corruption),
n.d.-b. As of 23 February 2016:
, Etat des signatures et ratifications du trait 191 (Protocole additionnel la Convention
pnale sur la corruption), n.d.-c. As of 23 February 2016:
, The European Commission for the Efficiency of Justice, n.d.-d. As of 23 February 2016:
http://www.coe.int/T/dghl/cooperation/cepej/default_en.asp
, Group of States against corruption (GRECO): The Fight against Corruption: a Priority
for the Council of Europe, n.d.-e. As of 23 February 2016:
PE 579.319
121
http://www.coe.int/t/dghl/monitoring/greco/general/1.%20The%20Fight%20against%20Co
rruption%20-%20A%20Priority%20for%20the%20CoE_en.asp
, How does GRECO work? n.d.-f. As of 23 February 2016:
http://www.coe.int/t/dghl/monitoring/greco/general/4.%20How%20does%20GRECO%20w
ork_en.asp
Collier, Paul, Martina Kirchberger, and Mns Sderbom. 2015. The cost of road infrastructure
in low and middle income countries. The World Bank Economic Review. lhv037.
Corruption Watch UK. 2015. Corruption Watch submission to Cabinet Office consultation on UK
transposition of new EU Procurement Directives. As of 23 February 2016
http://www.cw-uk.org/wp-content/uploads/2015/03/Corruption-Watch-submission-toCabinet-Office-consultation-on-UK-transposition-of-new-EU-Procurement-Directives.docx
Council of the EU. 1995. Act of 26 July 1995 Drawing up the Convention on the Protection of the
European Communities Financial Interests. OJ C 316 27 November 1995.
, Act of 27 September 1996 drawing up a Protocol to the Convention on the Protection of
the European Communities Financial Interests, 1996. OJ C 313 23 October 1996.
, Act of 19 June 1997 drawing up the Second Protocol of the Convention on the protection
of the European Communities financial interests, 1997. OJ C 221, 19 July 1997.
, Council Decision of 6 April 2009 establishing the European Police Office (Europol).
2009a. 2009/371/JHA
, The Stockholm Programme An open and secure Europe serving and protecting the
citizens. 2009b. 17024/09 CO EUR-PREP 3 JAI 896 POLGEN 229
, Plenary Session of the European Parliament, 11 September 2012: Council and
Commission statements - Political situation in Romania, 2012. 13780/12 PE 413 INST 528
, Protecting the Unions financial interests by means of criminal law, 2013. 10232/13
DROIPEN 67 JAI 439 GAF 25 FIN 303 CADREFIN 126 CODEC 1260.
, EU priorities for cooperation with the Council of Europe in 2016-2017, 2015a. 14919/15
COSCE 7 CFSP/PESC 831 COHOM 121.
, Proposal for a Council Regulation on the establishment of the European Public
Prosecutor's Office. 2015. Report on the State of Play. 15100/15 EPPO 48 EUROJUST 202
CATS 135 FIN 881 COPEN 345 GAF 54. As of 23 February 2016:
http://db.eurocrim.org/db/en/doc/2404.pdf
, Proposal for a Regulation on the establishment of the European Public Prosecutor's
Office - Report from the EPPO conference on 16-17 April 2015 at ERA. 2015c. 10283/15
EPPO 32 EUROJUST 129 CATS 67 FIN 464 COPEN 171 GAF 20.
PE 579.319
122
Coviello, Decio, and Mario Mariniello. 2014. Publicity requirements in public procurement:
Evidence from a regression discontinuity design. Journal of public economics, Vol. 109. 76100.
Davies, A. 2013. Initial appraisal of a European Commission Impact Assessment: European Commission
proposal for a Council Regulation on the establishment of the European Public Prosecutors Office.
EPRS, Ex-Ante Impact Assessment. PE 514.087. December 2013.
De Capitani, E. 2015. The new Europol: no more European FBI, not yet European NSA, in
FREE Group: European Area of Freedom Security & Justice. 6 December 2015, 2015. As of 23
February 2016:
http://free-group.eu/2015/12/06/the-new-europol-no-more-european-fbi-not-yet-europeannsa/
Dimitrov, G., K. Haralampiev, and Georgieva Z. 2014. Problemut s biorbata sreshtu korupciyata v
Bulgaria, vidyan prez dokladite po mehanizma za sutrudnichestvo I proverka v perioda 2007-2012.
Sofia.
Dimitrov, G., K. Haralampiev, S. Stoychev, and L. Toneva-Metodieva. 2014. The Cooperation
and Verification Mechanism: Shared Political Irresponsibility between the European
Commission and the Bulgarian Governments. Sofia: St. Kliment Ohridski University
Press.
Dimitrova, Antoaneta. 2015. The Effectiveness and Limitations of Political Integration in
Central and Eastern European Member States: Lessons from Bulgaria and Romania.
Maximizing the integration capacity of the European Union: Lessons and prospects for enlargement
and beyond (maxcap project), Working Paper. No. 10.
Dimitrova, Antoaneta, and Aron Buzogny. 2014. PostAccession PolicyMaking in Bulgaria
and Romania: Can Nonstate Actors Use EU Rules to Promote Better Governance? JCMS:
Journal of Common Market Studies, Vol. 52, No. 1. 139-156.
Disch, Arne, Endre Vigeland, and Geir Sundet. 2009. Anti-corruption approaches: A literature
review. Study. Vol. 2. 2208.
Doig, Alan, and Michael Levi. 2009. Inter-agency work and the UK public sector investigation
of fraud, 19962006: joined-up rhetoric and disjointed reality. Policing & Society. Vol. 19.
No. 3. 199-215.
, A case of arrested development? Delivering the UK National Fraud Strategy within
competing policing policy priorities. Public Money & Management, Vol. 33, No. 2, 2013. 145152.
Dolan, C. 2014. We have an anti-corruption report: So now what? Transparency International. 3
February 2014. As of 23 February 2016:
http://www.transparencyinternational.eu/2014/02/we-have-an-eu-anti-corruption-report-sonow-what/
Dori, Adriani. 2015. The EU Justice Scoreboard Judicial evaluation as a new governance tool,
MPILux Working Paper.
PE 579.319
123
Douglas-Scott, S. 2014. Opinion 2/13 on EU accession to the ECHR: a Christmas bombshell

from the European Court of Justice. Verfassungsblog, 24 December 2014. As of 23 February
2016:
http://verfassungsblog.de/en/opinion-213-eu-accession-echr-christmas-bombshell-europeancourt-justice/
Doyle, Charles. 2009. Qui Tam: The False Claims Act and Related Federal Statutes.
Easterly, William, and Ross Levine. 1997. Africa's growth tragedy: policies and ethnic
divisions. The Quarterly Journal of Economics. 1203-1250.
EC, Communication from the Commission to the Council, the European Parliament and the
European Economic and Social Committee on a comprehensive EU policy against
corruption. 2003.
, Eurobarometer 245: Organised crime and corruption. 2006.
, Eurobarometer 291: The attitudes of Europeans towards corruption. 2008a.
, Report from the Commission to the European Parliament and the Council on Progress
in Romania/Bulgaria under the Cooperation and Verification Mechanism. 2008b.
, Eurobarometer 325: Attitude of Europeans towards corruption. 2009.
, Communication from the Commission: Europe 2020: A strategy for smart, sustainable
and inclusive growth. 2010.
, Commission Staff Working Document: Annual Public Procurement Implementation
Review 2012. 2011a.
, Commission Staff Working Document: Impact Assessment of an Initiative on
Concessions. 2011b.
, Commission Staff Working Paper: Evaluation Report: Impact and Effectiveness of EU
Public Procurement Legislation. 2011c.
, Communication from the Commission to the Parilament, the Council and the European
Economic and Social Committee: Fighting Corruption in the EU. 2011d.
, Communication from the Commission to the European Parliament, the Council and the
European Economic and Social Committee: Participation of the European Union in the
Council of Europe Group of States against Corruption (GRECO). 2012a.
, Eurobarometer 374: Corruption. 2012b.
, Proposal for a Directive of the European Parliament and of the Council on the fight
against fraud to the Union's financial interests by means of criminal law. 2012c.
PE 579.319
124
, Commission Staff Working Document: Impact Assessment Accompanying the Proposal

for a Council Regulation on the Establishment of the European Public Prosecutors Office.
2013a.
, Communication from the Commission to the European Parliament, the Council, the
European Economic and Social Committee and the Committee of the Regions. 2013b.
European Economic and Social Committee and the Committee of the Regions: Improving
OLAF's governance and reinforcing procedural safeguards in investigations:
A step-by-step approach to accompany the establishment of the European Public Prosecutor's
Office. 2013c.
, Meeting Report: 8th Meeting of the Group of Experts on Corruption, 17 September 2013,
Brussels. 2013d. As of 23 February 2016:
http://ec.europa.eu/transparency/regexpert/index.cfm?do=groupDetail.groupDetailDoc&id=
9709&no=1
, Proposal for a Council Regulation on the establishment of the European Public
Prosecutor's Office. 2013e.
, Protecting taxpayers money against fraud: Commission proposes European Public
Prosecutor's Office and reinforces OLAF procedural guarantees. 2013f. As of 23 February
2016:
http://europa.eu/rapid/press-release_IP-13-709_en.htm
, Communication from the Commission to the European Parliament and the Council: A
new EU Framework to strenghten the Rule of Law 2014a.
European Central Bank, the European Economic and Social Committee, the Committee of
the Regions and the European Investment Bank: Annual Growth Survey 2015. 2014b.
European Economic and Social Committee and the Committee of the Regions. 2014c.
, EU Anticorruption Report: Annex: Ireland. 2014d.
, Eurobarometer 374: Businesses attitudes towards corruption in the EU. 2014e.
, Eurobarometer 397: Corruption, February 2014. 2014f. As of 23 February 2016:
http://ec.europa.eu/public_opinion/archives/ebs/ebs_397_en.pdf
, A new era for EU Justice and Home Affairs policies. 2014g. As of 23 February 2016:
http://europa.eu/rapid/press-release_IP-14-2266_en.htm
PE 579.319
125
, Report from the Commission to the Council and the European Parliament: EU
Anticorruption Report. 2014h. As of 23 February 2016:
http://ec.europa.eu/dgs/home-affairs/e-library/documents/policies/organized-crime-andhuman-trafficking/corruption/docs/acr_2014_en.pdf
European Central Bank, the European Economic and Social Committee, the Committee of
the Regions and the European Investment Bank: Annual Growth Survey 2016. 2015a.
, Communication from the Commission: Updating of data used to calculate lump sum
and penalty payments to be proposed by the Commission to the Court of Justice in
infringement proceedings. 2015b.
, Eurobarometer 428: Businesses attitudes towards corruption in the EU. 2015c.
, Report from the Commission: Monitoring the application of Union Law: 2014 Annual
Report. 2015d.
, TED Structured Dataset (2009-2014), Tenders Electronic Daily, supplement to the
Official Journal of the European Union. Brussels. 2015e.
, Rule of Law in Poland: Commission starts dialogue. 2016. As of 23 February 2016:
http://ec.europa.eu/news/2016/01/20160113_en.htm
, Anti-corruption experience sharing programme. n.d.-a. As of 23 February 2016:
http://ec.europa.eu/dgs/home-affairs/what-we-do/policies/organized-crime-and-humantrafficking/corruption/experience-sharing-programme/index_en.htm
, Assessing ongoing progress by Bulgaria and Romania. n.d.-b. As of 23 February 2016:
http://ec.europa.eu/cvm/index_en.htm
, Competition: Leniency. n.d.-c. As of 23 February 2016:
http://ec.europa.eu/competition/cartels/leniency/leniency.html
, EU Justice scoreboard. n.d.-d. As of 23 February 2016:
http://ec.europa.eu/justice/effective-justice/scoreboard/index_en.htm
, European Commission at work: Infringement procedure. n.d.-e. As of 23 February
2016:
http://ec.europa.eu/atwork/applying-eu-law/infringements-proceedings/index_en.htm
, Protecting EU interests: The Early Detection and Exclusion System: EDES. n.d.-f. As of
23 February 2016:
PE 579.319
126
http://ec.europa.eu/budget/explained/management/protecting/protect_en.cfm
, Public Procurement. n.d.-g. As of 23 February 2016:
http://ec.europa.eu/internal_market/scoreboard/performance_per_policy_area/public_procu
rement/
, Register of Commission Expert Groups: Expert Group on Corruption (E02725), n.d.-h.
As of 23 February 2016:
http://ec.europa.eu/transparency/regexpert/index.cfm?do=groupDetail.groupDetail&groupI
D=2725
, The reports on progress in Bulgaria and Romania. n.d.-i. As of 23 February 2016:
http://ec.europa.eu/cvm/progress_reports_en.htm
, The reports on progress in Bulgaria and Romania. n.d.-j. As of 23 February 2016:
http://ec.europa.eu/cvm/progress_reports_en.htm
, Six-pack? Two-pack? Fiscal compact? A short guide to the new EU fiscal governance.
n.d.-k. As of 23 February 2016:
http://ec.europa.eu/economy_finance/articles/governance/2012-03-14_six_pack_en.htm
Eckes, Christina, and Theodore Konstadinides. 2011. Crime within the area of freedom, security and
justice: a European public order: Cambridge University Press.
Ehlrich, Isaac, and Francis T Lui. 1999. Bureaucratic corruption and endogenous growth. J Polit
Econ. Vol. 107. No. 6. 270-293.
EP. 2013. European Parliament resolution of 23 October 2013 on organised crime, corruption
and money laundering: recommendations on action and initiatives to be taken.
, Answer given by Mr emeta on behalf of the Commission. 2014a.
, European Parliament resolution of 12 March 2014 on evaluation of justice in relation to
criminal justice and the rule of law 2014b.
, Report on the situation of fundamental rights in the European Union (2012). 2014c.
, Draft interim report on the proposal for a Council regulation on the establishment of the
European Public Prosecutors Office. 2015a.
, European Parliament resolution of 29 April 2015 on the proposal for a Council
regulation on the establishment of the European Public Prosecutors Office. 2015b.
, Europol: deal on new powers to step up EU police cooperation and fight terrorism.
2015c. As of 23 February 2016:
PE 579.319
127
http://www.europarl.europa.eu/news/en/news-room/20151130IPR05456/Europol-deal-onnew-powers-to-step-up-EU-police-cooperation-and-fight-terrorism
, Fight against corruption and follow-up of the CRIM committee resolution. 2015d.
, Luxleaks whistleblower Deltour wins broad support in Parliament. 2015e. As of 23
February 2016:
http://www.europarl.europa.eu/news/en/news-room/20150601IPR61309/Luxleakswhistleblower-Deltour-wins-broad-support-in-Parliament
, Question for written answer to the Commission. 2015f.
, Question for written answer to the Commission. 2015g.
, Reply to a question. 2015h.
, Rules of Procedure of the European Parliament: ANNEX VI : Powers and
responsibilities of standing committees. 2015i. As of 23 February 2016:
http://www.europarl.europa.eu/sides/getLastRules.do?language=en&reference=RESP-LIBE
, Setting up the European Public Prosecutor's Office. 2015j. As of 23 February 2016:
http://www.europarl.europa.eu/EPRS/EPRS-AaG-554196-Setting-up-European-PublicProsecutors-Office-FINAL.pdf
, Transparency, accountability and integrity in the EU institutions. Committee on
Constitutional Affairs Draft Report. 2015/2041(INI). 2015k.
, Nothing but the truth: The protection of whistleblowers at EU level. ALDE Seminar 27
January 2016. As of 23 Feburary 2016:
https://alde.livecasts.eu/nothing-but-the-truth/about
EP and Council of the EU. 1999. Regulation No. 1073/1999 of 25 May 1999 concerning
investigations conducted by the European Anti-Fraud Office (OLAF).
Esarey, Justin. 2015. Using Interaction Terms as Instrumental Variables for Causal
Identification: Does Corruption Harm Economic Development?
Esarey, Justin, and Gina Chirillo. 2013. Fairer Sex or Purity Myth? Corruption, Gender, and
Institutional Context, Politics & Gender. Vol. 9. No. 04. 361-389.
EU. 2002. Treaty on European Union (Consolidated Version), Treaty of Maastricht 1992, code
edition dated 24 December 2002.
, Directive 2004/17/EC of the European Parliament and of the Council of 31 March 2004
coordinating the procurement procedures of entities operating in the water, energy,
transport and postal services sectors. 2004a.
PE 579.319
128
, Directives 2004/18/EC of the European Parliament and of the Council of 31 March 2004
on the coordination of procedures for the award of public works contracts, public supply
contracts and public service contracts. 2004b.
, Consolidated version of the Treaty on the Functioning of the European Union. 2007.
, Directive 2014/23/EU of the European Parliament and of the Council of 26 February
2014 on the award of concession contracts. 2014a.
2014 on public procurement and repealing Directive 2004/18/EC. 2014b.
2014 on procurement by entities operating in the water, energy, transport and postal
services sectors and repealing Directives 2004/17/EC and 2004/18/EC, 2014c.
, Directive 2014/42/EU of the European Parliament and of the Council of 3 April 2014 on
the freezing and confiscation of instrumentalities and proceeds of crime in the European
Union. 2014d.
, Directive 2014/55/EU of the European Parliament and of the Council of 16 April 2014
on electronic invoicing in public procurement. 2014e.
, Directive 2014/95/EU of the European Parliamentand of the Council of 22 October 2014
amending Directive 2013/34/EU as regards disclosure of non-financial and diversity
information by certain large undertakings and groups. 2014f.
, Directive 2015/849/EU of the European Parliament and of the Council of 20 May 2015
on the prevention of the use of the financial system for the purposes of money laundering
or terrorist financing, amending Regulation (EU) No 648/2012 of the European Parliament
and of the Council, and repealing Directive 2005/60/EC of the European Parliament and of
the Council and Commission Directive 2006/70/EC, 2015a.
, Regulation (EU, Euratom) 2015/1929 of the European Parliament and of the Council of
28 October 2015 amending Regulation (EU, Euratom) No 966/2012 on the financial rules
applicable to the general budget of the Union. 2015b.
EUR-lex, Procedure 2012/0193/COD, n.d. As of 23 Feburary 2016:
http://eur-lex.europa.eu/procedure/EN/201814
European Ombudsman, Decision of the European Ombudsman closing her own-initiative
inquiry OI/1/2014/PMC concerning whistleblowing. 2014. As of 23 February 2016:
http://www.ombudsman.europa.eu/en/cases/decision.faces/en/59114/html.bookmark
European Univerity Institute, Whistleblowing Protection Laws EU Member States Laws. n.d.
As of 23 February 2016:
http://journalism.cmpf.eui.eu/maps/whistleblowing/
PE 579.319
129
Europol. 2013. SOCTA 2013: EU Serious and Oganised Crime Threat Assessment. As of 23 February
2016:
https://www.europol.europa.eu/sites/default/files/publications/socta2013.pdf
, SIENA. n.d. As of 23 February 2016:
https://www.europol.europa.eu/content/page/siena-1849
Fazekas, M., and G. Kocsis. 2015. Uncovering High-Level Corruption: Cross-National
Corruption Proxies Using Government Contracting Data. Available at SSRN 2711932.
Fazekas, M., and B. Toth. 2016. Infrastructure for whom? Corruption risks in infrastructure
provision across Europe, in Wegrich K., G. Hammerschmid, and Kostka G., eds.,
Governance Report 2016, Oxford: Oxford University Press.
Fazekas, Mihly, Jana Chvalkovska, Jiri Skuhrovec, Istvn Jnos Tth, and Lawrence P King.
2013. Are EU funds a corruption risk? The impact of EU funds on grand corruption in
Central and Eastern Europe. The Anticorruption Frontline. The ANTICORRP Project, Vol. 2.
68-89.
Fjeldstad, Odd-Helge, and Jan Isaksen. 2008. Anti-Corruption Reforms: Challenges, Effects and
Limits of World Bank Support. Background Paper to Public Sector Reform: What Works
and Why? An IEG Evaluation of World Bank Support. An IEG Evaluation of World Bank
Support. IEG Working Paper. Vol. 7. 92.
Foellmi, Reto, and Manuel Oechslin. 2007. Who gains from non-collusive corruption? Journal of
development Economics. Vol. 82. No. 1. 95-119.
FRA, EU Charter of Fundamental Rights: Article 51 - Field of application. n.d.-a. As of 23
February 2016:
http://fra.europa.eu/en/charterpedia/article/51-field-application
, Fundamental rights indicators. n.d.-b. As of 23 February 2016:
http://fra.europa.eu/en/project/2011/fundamental-rights-indicators
Franssen, V. 2013. Proposed Regulation on the European Public Prosecutor, in European Law
Blog. 8 August 2013. As of 23 February 2016:
http://europeanlawblog.eu/?p=1887
G20 Australia, 2015-16 G20 Anti-Corruption Action Plan. 2014a. As of 23 February 2016:
http://www.g20australia.org/sites/default/files/g20_resources/library/201516%20_g20_anti-corruption_action_plan_0.pdf
, 2015-16 G20 Anti-Corruption Implementation Plan, 2014b. As of 23 February 2016:
PE 579.319
130
http://www.g20australia.org/sites/default/files/g20_resources/library/G20_anti_corruption
_working_group_implementation_plan_2015_16.docx
Garcia-Murillo, Martha. 2013. Does a government web presence reduce perceptions of
corruption? Information Technology for Development. Vol. 19. No. 2. 151-175.
Gates, M.N. 2011. The Impact of Sarbanes-Oxley on Private Companies and Corporate Governance Best
Practices. Practical Law Company.
Gateva, Eli. 2010. Post-Accession Conditionality-Support Instrument for Continuous Pressure? Free
University Berlin.
Glaeser, Edward L, Rafael La Porta, Florencio Lopez-de-Silanes, and Andrei Shleifer. 2004. Do
institutions cause growth? Journal of economic Growth. Vol. 9 No. 3. 271-303.
Government Transparency Institute. Homepage. (n.d.). As of 23 February 2016:
http://www.govtransparency.eu/
Gundlach, Erich, and Martin Paldam. 2009. The transition of corruption: From poverty to
honesty. Economics Letters. Vol. 103. No. 3. 146-148.
Gupta, Sanjeev, Hamid Davoodi, and Rosa Alonso-Terme. 2002. Does corruption affect income
inequality and poverty? Economics of governance. Vol. 3. No. 1. 23-45.
Hall, Robert E, and Charles I Jones. 1999. Why do some countries produce so much more output per
worker than others? National bureau of economic research.
Hanna, Rema, S Bishop, K Durlacher, S Nadel, and G Scheffler. 2010. The effectiveness of anticorruption policy: what has worked, what hasnt, and what we dont knowa systematic
review protocol. DFID Systematic Review.
Hawken, Angela, and Gerardo L Munck. 2009. Do you know your data? Measurement validity
in corruption research. Unpublished typescript, Pepperdine University and University of
Southern California, Malibu, CA, and Los Angeles.
Hawley, S. 2015. Yawn and Youll Miss It: New EU Debarment Rules, Old Problems. As of 23
February 2016:
http://www.cw-uk.org/2015/03/04/yawn-and-youll-miss-it-new-eu-debarment-rules-oldproblems/
Herzfeld, Thomas, and Christoph Weiss. 2003. Corruption and legal (in) effectiveness: an
empirical investigation. European Journal of Political Economy. Vol. 19. No. 3. 621-632.
Heywood, Paul M, and Jonathan Rose. 2014. Close but no Cigar: the measurement of
corruption." Journal of Public Policy. Vol. 34. No. 03. 507-529.
Inman, Kris, and Josephine Andrews. 2010. Political participation in Africa: Evidence from
survey and experimental research. Annual Meeting of the Midwest Political Science
Association. 22-25.
PE 579.319
131
Institute for Public Policy. 2010. Effective Mechanisms and Practices for Fighting Corruption in the
Black Sea Region: Lessons Learnt and Further Steps for Romania, Bulgaria, Turkey and Georgia.
Bucharest: The Black Sea Trust for Regional Cooperation.
Ivanova, I. 2009. The Case of Bulgaria. in Alegre. Edited by S., I. Ivanova, and D. Denis-Smith,.
Safeguarding the Rule of Law in an Enlarged EU The Cases of Bulgaria and Romania: Centre for
European Policy Studies.
Jackson Lewis Corporate Governance Practice Group. 2009. Compliance & Ethics Requirements for
Federal Contractors, Jackson Lewis LLP Special Report.
Johnson, Joel W, and Jessica S Wallack. 1997. Electoral systems and the personal vote. Database.
Kaufmann, Daniel. 2005. Myths and realities of governance and corruption. Available at SSRN
829244.
Kochenov, Dimitry, and Laurent Pech. 2015. Upholding the Rule of Law in the EU: On the
Commission's Pre-Article 7 Procedureas a Timid Step in the Right Direction. Robert
Schuman Centre for Advanced Studies Research Paper. No. 2015/24.
Kostadinova, Tatiana. 2009. Abstain or rebel: Corruption perceptions and voting in East
European elections. Politics & Policy. Vol. 37. No. 4. 2009. 691-714.
Kurtz, Marcus J, and Andrew Schrank. 2007. Growth and governance: Models, measures, and
mechanisms. Journal of politics. Vol. 69. No. 2. 538-554.
Lambsdorff, Johann Graf. 2007. The institutional economics of corruption and reform: theory, evidence
and policy: Cambridge University Press.
Linklaters. 2014. Financial Crime Update. 29 May 2014, 2014. As of 23 February 2016:
http://www.linklaters.com/Insights/Financial-Crime-Update/Financial-Crime-Update-May2014/Pages/Globalnews.aspx
Lord, Nicholas, and Michael Levi. 2015. Determining the adequate enforcement of white-collar
and corporate crimes in Europe. in Erp. Edited by J., W. Huisman, and G. Vande Walle.
The Routledge Handbook of White-Collar and Corporate Crime in Europe. 39-56.
Manacorda, S., and C. Grasso. 2014. Debarment of Companies in the European Union, ROXIN
Alliance. As of 23 February 2016:
http://www.roxin-alliance.org/uploads/media/debarment_in_europe.pdf
Mauro, Paolo. 195. Corruption and growth. The quarterly journal of economics. 681-712.
Melgar, Natalia, Mximo Rossi, and Tom W Smith. 2010. The perception of corruption.
International Journal of Public Opinion Research. Vol. 22. No. 1. 120-131.
Meyer-Sahling, Jan-Hinrik. 2009. Sustainability of civil service reforms in Central and Eastern
Europe five years after EU accession.
PE 579.319
132
Minister of Foreign Affairs of the Netherlands. 2015. Letter to the house of Representatives about
substantive preparations for the 2016 Dutch Presidency of the Council of the European Union. As
of 23 February 2016:
https://www.government.nl/topics/european-union/documents/parliamentarydocuments/2015/01/23/letter-to-the-house-of-representatives-about-substantivepreparations-for-the-2016-dutch-presidency-of-the-council-of-the-europ
Mistry, Jamshed J, and Abu Jalal. 2012. An empirical analysis of the relationship between egovernment and corruption. The International journal of digital accounting research. Vol. 12.
No. 18. 145-176.
Mo, Pak Hung. 2001. Corruption and economic growth. Journal of comparative economics, Vol.
29. No. 1. 66-79.
Montinola, Gabriella R, and Robert W Jackman. 2002. Sources of corruption: a cross-country
study. British Journal of Political Science. Vol. 32. No. 01. 147-170.
Mller, JanWerner. 2015. Should the EU Protect Democracy and the Rule of Law inside
Member States? European Law Journal. Vol. 21. No. 2. 141-160.
Mungiu-Pippidi, Alina. 2006. Corruption: Diagnosis and treatment. Journal of democracy. Vol.
17. No. 3. 86-99.
, The good, the bad and the ugly: controlling corruption in the European union.
Advanced Policy Paper for Discussion in the European Parliament. Vol. 9. 2013.
Neupane, Arjun, Jeffrey Soar, Kishor Vaidya, and Jianming Yong. 2012. Role of public eprocurement technology to reduce corruption in government procurement. Proceedings of
the 5th International Public Procurement Conference (IPPC5). 304-334.
OECD. 1997. Convention on Combatting Bribery of Foreign Public Officials in International Business
Transactions: OECD.
, Recommendation of the Council on Improving Ethical Conduct in the Public Service
Including Principles for Managing Ethics in the Public Service. 1998.
, Managing Conflict of Interest in the Public Service: OECD Guidelines and Country
Experiences. Paris: OECD. 2003.
, Boosting Integrity, Fighting Corruption: CleanGovBiz Integrity in Practice. 2012. As of 23
February 2016:
http://www.oecd.org/daf/anti-bribery/50350066.pdf
, Government at a Glance 2013. Paris. 2013a.
, Issues Paper on Corruption and Growth. 2013b. As of 23 February 2016:
http://www.oecd.org/g20/topics/anti-corruption/Issue-Paper-Corruption-and-EconomicGrowth.pdf
PE 579.319
133
, OECD Convention on Combatting Bribery of Foreign Public Officials in International Business

Transactions: Ratification Status as of 21 May 2014. 2014a.
, The rationale for fighting corruption: CleanGovBiz Integrity in Practice. 2014b. As of 23
February 2016:
http://www.oecd.org/cleangovbiz/49693613.pdf
, Country monitoring of the OECD Anti-Bribery Convention. n.d.-a. As of 23 February
2016:
http://www.oecd.org/daf/anti-bribery/countrymonitoringoftheoecdantibriberyconvention.htm
, OECD Convention on Combatting Bribery of Foreign Public Officials in International
Business Transactions. n.d.-b. As of 23 February 2016:
http://www.oecd.org/corruption/oecdantibriberyconvention.htm
Oechslin, Manuel, and Reto Foellmi. 2003. Who Gains from Non-Collusive Corruption? Royal
Economic Society Annual Conference 2003.
OLAF. 2011. The OLAF Report 2011, 2011. As of 23 February 2016:
http://ec.europa.eu/anti_fraud/documents/reports-olaf/2011/olaf_report_2011_en.pdf
, The OLAF Report 2014. 2014. As of 23 February 2016:
http://ec.europa.eu/anti_fraud/documents/reports-olaf/2011/olaf_report_2011_en.pdf
Olken, Benjamin A, and Patrick Barron. 2007. The simple economics of extortion: evidence from
trucking in Aceh. National Bureau of Economic Research Cambridge, Mass. USA.
Palokaj, A. 2010. No control mechanism for Croatia after EU entry. EUObserver. 16 November
2010.
Paulhus, Delroy L. 1984. Two-component models of socially desirable responding. Journal of
Personality and Social Psychology. Vol. 46. No. 3. 598.
Pech, Laurent. 2015. The EU as a global rule of law promoter: the consistency and effectiveness
challenges. Asia Europe Journal. 1-18.
Pellegrini, Lorenzo, and Reyer Gerlagh. 2004. Corruption's effect on growth and its
transmission channels. Kyklos. Vol. 57. No. 3. 429-456.
Podsakoff, Philip M, Scott B MacKenzie, Jeong-Yeon Lee, and Nathan P Podsakoff. 2003.
Common method biases in behavioral research: a critical review of the literature and
recommended remedies. Journal of applied psychology. Vol. 88. No. 5. 879.
Pollitt, Christopher. Moderation in all things: international comparisons of governance quality.
Financial Accountability & Management. Vol. 27. No. 4. 437-457.
PE 579.319
134
Pop, V. 2011. EU to monitor anti-corruption measures in Member States. EUObserver, 6 June

2011. As of 23 February 2016:
https://euobserver.com/justice/32444
PricewaterhouseCoopers, and i-Force. 2011. Corruption and conflict of interest in the European
Institutions: the effectiveness of whistleblowers. Prepared for the European Parliament's
Committee on Budgetary Control (COCOBU), IP/D/ALL/FWC/2009-056
PricewaterhouseCoopers and Ecorys. 2013. Identifying and reducing corruption in public
procurement in the EU.
Razafindrakoto, Mireille, and Francois Roubaud. 2006. Are international databases on
corruption reliable? A comparison of expert opinion surveys and household surveys in
sub-Saharan Africa.' Manuscript. IRD/DIAL. As of 25 February 2016:
http://www.dial.ird.fr/media/ird-sites-d-unites-derecherche/dial/documents/publications/doc_travail/2006/2006-17
Restarting The Future. Homepage. n.d. As of 23 February 2016:
http://www.restartingthefuture.eu/about
Riley, A. 2013. Croatia and the EU. New York Times. 28 June 2013.
Rizzo, G. 2015 (EPPO) European Public Prosecutor: also the European Parliament wants a
say in European Area of Freedom Security & Justice: Free Group... 27 April 2015. As of 23
February 2016:
http://free-group.eu/2015/04/27/eppo-european-public-prosecutor-also-the-europeanparliament-wants-a-say/
Rohde-Liebenau, B. 2006. Whistleblowing Rules: Best Practice; Assessment and Revision of Rules
Existing in EU Institutions., European Parliament, Directorate General Internal Policies of
the Union, Budgetary Support Unit, Budgetary Affairs, IPOL/D/CONT/ST/2005_58.
Rohwer, Anja. 2009. Measuring Corruption: A comparison between the transparency
international's corruption perceptions index and the world bank's worldwide governance
indicators. CESifo DICE Report, Vol. 7. No. 3. 42-52.
Rothstein, B, and SREN Holmberg. 2011. Correlates of Corruption. University of Gothenburg,
Quality of Government Institute Working Paper. No. 2011. 12.
Rubin, J., J. Taylor, A. Pollitt, J. Krapels, and M. Pardal. 2014. Intolerance in Western Europe.
2014.
Santoro, D., M. Kumar, M. Turola, L. Ferrante, and F. Quarta. 2014. Blowing the Whistle on
Corruption. Campaign for an European Directive in Defence of Whistleblowers.
PE 579.319
135
Sedelmeier, Ulrich. 2014. Anchoring democracy from above? The European Union and
democratic backsliding in Hungary and Romania after accession. JCMS: Journal of Common
Market Studies. Vol. 52. No. 1. 105-121.
Seligson, Mitchell A. 2006. The measurement and impact of corruption victimization: Survey
evidence from Latin America. World Development. Vol. 34. No. 2. 381-404.
Shimabukuro, Jon O, and L Paige Whitaker. 2012. Whistleblower Protections Under Federal Law:
An Overview. 7-5700.
Stephenson, Paul, and Michael Levi. 2012. The Protection of Whistleblowers: A study on the
feasibility of a legal instrument on the protection of employees who make disclosures in the public
interest. As of 23 February 2016:
http://www.coe.int/t/dghl/standardsetting/cdcj/Whistleblowers/CDCJ(2012)9E_Final.pdf
Stockemer, Daniel, Bernadette LaMontagne, and Lyle Scruggs. 2013. Bribes and ballots: The
impact of corruption on voter turnout in democracies. International Political Science Review.
Vol. 34. No. 1. 74-90.
Sundstrm, Aksel, and Daniel Stockemer. 2013. Quality of government affect voter turnout in
the European regions. QoG Working Paper Series. Vol. 2013. No. 6: 6.
Svensson, Jakob. 2005. Eight questions about corruption. The Journal of Economic Perspectives.
Vol. 19. No. 3. 19-42.
Tavits, Margit. 2008. Representation, corruption, and subjective well-being. Comparative
Political Studies. Vol. 41. No. 12. 1607-1630.
Thomas, Melissa A. 2010. What do the worldwide governance indicators measure&quest.
European Journal of Development Research. Vol. 22. No. 1. 31-54.
Thomas, W., and A. Maguire. 2014 Changes Ahead for Corporate Sustainability Disclosures
Following Publication of New EU Directive. Wilkie Farr & Galagher LLP. Client
Memorandum. As of 23 February 2016:
http://www.willkie.com/~/media/Files/Publications/2014/12/Changes_Ahead_for_Corpor
ate_Sustainability_Disclosures_Following_Publication_of_New_EU_Directive.pdf
Tourangeau, Roger, and Tom W Smith. 1996. Asking sensitive questions the impact of data
collection mode, question format, and question context. Public opinion quarterly. Vol. 60,
No. 2. 1996: 275-304.
Transparency International. 2005. Global Corruption Report 2005. Berlin: Transparency
International.
, Controlling Corporate Lobbying and Financing of Political Activities. Policy Position
06/2009. 2009.
, Combatting Corruption in the EU through E-Procurement. 2011. As of 23 February
2016:
PE 579.319
136
https://blog.transparency.org/2011/02/16/combatting-corruption-in-the-eu-through-eprocurement/
, Money, Politics, Power: Corruption Risks in Europe. June 2012. As of 23 February 2016:
https://www.transparency.org/enis/report
, UN Convention against Corruption: Progress Report. 2013. As of 23 February 2016:
http://www.transparency.org/whatwedo/pub/un_convention_against_corruption_progress_
report_2013
, 11 years after signing, German Parliament approves Global Corruption Agreement.
2014a. As of 23 February 2016:
http://www.transparency.org/news/pressrelease/11_years_after_signing_german_parliament
_approves_global_corruption_agreeme
, The EU Integrity System. 2014b. As of 23 February 2016:
http://www.transparencyinternational.eu/european-union-integrity-system-study/the-euisreport-latest-news/
, EU institutions drag their feet and still fail to protect whistleblowers. 2015a. As of 23
Feburary 2016:
http://www.transparencyinternational.eu/2015/03/eu-institutions-drag-their-feet-and-stillfail-to-protect-whistleblowers/
, Lobbying in Europe: Hidden influence, privileged access. 2015b. As of 23 February 2016:
http://www.transparency.org/news/feature/europe_a_playground_for_special_interests_ami
d_lax_lobbying_rules.
, Corruption by country/territory. n.d. As of 23 February 2016:
https://www.transparency.org/country/
Treisman, Daniel. 2000. The causes of corruption: a cross-national study. Journal of public
economics. Vol. 76. No. 3. 399-457.
, What have we learned about the causes of corruption from ten years of cross-national
empirical research? Annu. Rev. Polit. Sci. Vol. 10. 2007. 211-244.
U4 Anti-Corruption Resource Centre. Homepage. n.d. As of 23 February 2016:
http://www.u4.no/
Ugur, Mehmet. 2014. Corruption's Direct Effects on PerCapita Income Growth: A
MetaAnalysis. Journal of Economic Surveys. Vol. 28. No. 3. 2014. 472-490.
PE 579.319
137
UN. 2014. United Nations E-government Survey 2014: E-government for the Future We Want,
ST/ESA/PAD/SER.E/188. 2014. As of 23 February 2016:
https://publicadministration.un.org/egovkb/Portals/egovkb/Documents/un/2014Survey/E-Gov_Complete_Survey-2014.pdf
, United Nations Public Administration Country Studies. n.d. As of 23 February 2016:
https://publicadministration.un.org/egovkb/Data-Center
UNDP. 2008. A Users Guide to Measuring Corruption, Norway: UNDP Oslo Governance Centre.
University of Gothenburg. N.d. Quality of Government data set. As of 23 February 2016:
http://qog.pol.gu.se/data
UNODC. 2001. Global Programme Against Corruption: Anti-Corruption Tool Kit.
, UN Guide for Anti-Corruption Policies, Vienna. 2003.
, United Nations Convention Against Corruption. 2004. As of 23 February 2016:
https://www.unodc.org/documents/brussels/UN_Convention_Against_Corruption.pdf
, Mechanism for the Review of Implementation of the United Nations Convention against
CorruptionBasic Documents. 2011. As of 23 February 2016:
https://www.unodc.org/documents/treaties/UNCAC/Publications/ReviewMechanismBasicDocuments/Mechanism_for_the_Review_of_Implementation_-_Basic_Documents__E.pdf
, United Nations Convention against Corruption Signature and Ratification Status as of
1
December
2015.
n.d.
As
of
25
February
2016:
https://www.unodc.org/unodc/en/treaties/CAC/signatories.html
Utrecht University. 2008. Catalogue of promising practices in the field of integrity, anti-corruption and
administrative measures against organized crime in the EU. Commissioned by the Ministry of the
Interior and Kingdom Relations of the Netherlands. As of 23 February 2016:
http://www.integriteitoverheid.nl/fileadmin/BIOS/data/Publicaties/Downloads/EU_Catalo
gus.pdf
Vachudova, Milada Anna, and Aneta Spendzharova. 2012. EU's Cooperation and Verification
Mechanism: Fighting Corruption in Bulgaria and Romania After EU Accession. Swedish Institute
for European Policy Studies.
Van Dijk, Jan. 2007. Mafia markers: assessing organized crime and its impact upon societies.
Trends in organized crime. Vol. 10. No. 4. 39-56.
Van Stolk, CHRISTIAN, and Emil Tesliuc. 2010. Toolkit on Tackling Error, Fraud and
Corruption in Social Protection Programs. World Bank, Washington, DC.
PE 579.319
138
Westman, Daniel P. 2005. The Significance of the Sarbanes-Oxley Whistleblower Provisions.

Lab. Law. Vol. 21: 141.
Wolf, Sebastian. 2010. Assessing Eastern Europe's anti-corruption performance: views from the
Council of Europe, OECD, and Transparency International. Global crime. Vol. 11. No. 2: 99121.
World Bank. 2004. Korea's move to e-procurement. 90.
, FAQs: Fraud & Corruption. 2009a. As of 23 February 2016:
http://web.worldbank.org/WBSITE/EXTERNAL/EXTSITETOOLS/0,,contentMDK:20147620~
menuPK:344192~pagePK:98400~piPK:98424~theSitePK:95474,00.html
, Fraud and Corruption. Awareness Handbook. Washington, DC: World Bank. 2009b.
, Governance & Anti-Corruption. n.d.-a. As of 23 February 2016:
http://web.worldbank.org/WBSITE/EXTERNAL/WBI/EXTWBIGOVANTCOR/0,,menuPK:1
740542~pagePK:64168427~piPK:64168435~theSitePK:1740530,00.html
, Six Questions on the Cost of Corruption with World Bank Institute Global Governance
Director Daniel Kaufmann. n.d.-b. As of 23 February 2016:
http://web.worldbank.org/WBSITE/EXTERNAL/NEWS/0,,contentMDK:20190295~menuPK:
34457~pagePK:34370~piPK:34424~theSitePK:4607,00.html
World Economic Forum. N.d. Competitiveness Rankings. As of 23 February 2016:
http://reports.weforum.org/global-competitiveness-report-2014-2015/rankings/
Zimring, Franklin E, and David T Johnson. 2005. On the comparative study of corruption.
British Journal of Criminology. Vol. 45. No. 6. 793-809.
PE 579.319
139
Appendix A: Provisions of Article 82 and 83 TFEU and summary of

costs of corruption estimates
Table A-25. Summary of provisions of Article 82 and Article 83 of TFEU
Article, TFEU
Art. 82, para 1
Art. 82 para 2
Art. 83 para 1
Art. 83 para 2
Competence
Measures to promote cooperation between judicial
authorities of the Member State in relation to
proceedings in criminal matters
approximation competence in the field of criminal
procedure law
approximation competence in the field of substantial
criminal law
Approximation competence in the field of
substantial law for areas which have been subject to
harmonisation measures
Form of action
directives, regulations,
decisions and more
directives
directives
directives
TABLE A-2: overview of existing high-level estimates of the cost of corruption

World Economic
European
World Bank
Forum
Commission
Cost of Corruption (annual):
$1 trillion worldwide
5% global GDP or
$2.6 trillion
Methodology/Description:
Figure based on 2001-2002

worldwide surveys of
enterprises and estimation
of bribes paid by
household users of public
services (from the WBI
governance and anticorruption surveys). Does
not include the extent of
embezzlement of public
funds (from central and
local budgets), or from
theft (or misuse) of public
assets.
Figure probably
based on WB
estimate of $1
trillion and the
assumption that the
cost of bribery was
around 3.5-5% of
global GDP. The
WEF highlights that
estimates suggest
the cost of
corruption amounts
to more than 5% of
global GDP (US$2.6
trillion) with more
than US$1 trillion
paid in bribes each
year.
Caveats/Limitations:
Estimate only includes

bribes paid; not includes
the extent of
embezzlement of public
funds (from central and
local budgets), or from
theft (or misuse) of public
assets. Furthermore, no
indirect costs of
corruption included
the methodology
how estimate was
calculated is not
well documented148
120bn at EU-level
Figure is based on
estimates by
specialised
institutions and
bodies, such as the
International
Chamber of
Commerce,
Transparency
International, UN
Global Compact,
World Economic
Forum, Clean
Business is Good
Business, 2009,
which suggest that
corruption amounts
to 5% of GDP at
world level.
the methodology
how estimate was
calculated is not
well documented;
does not include
indirect effects and
magnitude of the
estimate has been
questioned by
recent research
(Mungiu-Pippidu
(2013)
See
the
discussion
on
Global
Anitcorruption
Blog
(GAB):
http://globalanticorruptionblog.com/2015/12/22/where-does-the-2-6-trillion-corruption-cost-estimatecome-from/ As of 23 February 2016.
148
PE 579.319
140
Appendix B: Details on our approach to literature review

The overarching approach to the literature review was based on a snowballing strategy. The
starting point for the identification of relevant sources was a set of key documents outlined in
the terms of reference, complemented by sources identified in consultation with the
commissioning team and our senior advisor.
As part of the review of the initial set of literature, bibliographies of these sources were
searched for additional relevant material. This step was subsequently repeated for newly
identified sources. In addition, the research team also reviewed materials suggested by expert
interviewees.
PE 579.319
141
Appendix C: Descriptive evidence corruption indicators and

the economic, social and political costs
Figure C-1: Corruption and level of output (GDP per capita) EU-28 Member State
Note: average values 1995-2014.
PE 579.319
142
Figure C-2: Corruption and the rule of law - EU-28 Member State
PE 579.319
143
Figure C-3: Corruption and level of output (rule of law) EU-28 Member State
PE 579.319
144
Appendix D: Control variables used in the empirical analysis to

assess costs of corruption
Table D-1: Variables used in estimating the economic costs of corruption
Type of cost
Outcome variable
log GDP per

capita (PPP,
constant 2005)
QoG (WDI)
Economic
growth genuine
investment
(genuine wealth
per capita)
PE 579.319
Secondary School Enrolment Rate
Source
control
variables
WDI
Life Expectancy at birth
WDI
Initial log GDP (1985)
WDI
Government expenditure (% GDP)
WDI
Trade openness (% GDP)
WDI
Added value manufacturing (% GDP)
WDI
Added value service (% GDP)
WDI
Income share 20% highest
WDI
Gross capital formation (% GDP)
WDI
Level of democracy (polity 2)
QoG
Freedom of press score (Freedom House)
QoG
Personalism (Johnson & Wallack)
QoG
Proportion women lower house (IPU)
QoG
Presidentialism
DPI
WDI
WDI
Initial log GDP (1985)
WDI
WDI
WDI
QoG
Freedom of press score (Freedom House)
QoG
QoG
QoG
Presidentialism
DPI
Source outcome
variable
based on
calculations by
Aidt (2010)
using data from
WDI ('adjusted
net savings')
Control variables
145
Table D-2: Variables used in estimating the social costs of corruption

Type of
cost
Outcome
variable
Inequality (Giniindex; Income

share 20%
highest)
Social
Rule of Law
Organised Crime
PE 579.319
Source outcome
variable
QoG (WDI)
World Bank WGI
World Economic
Forum
146
Source
control
variables
WDI
WDI
WDI
WDI
WDI
QoG
Freedom of press score (Freedom

House)
QoG
QoG
Presidentialism
DPI
WDI
WDI
WDI
WDI
WDI
QoG

House)
QoG
QoG
Presidentialism
DPI
WDI
WDI
QoG

House)
QoG
QoG
Presidentialism
DPI
log GDP per capita
WDI
WDI
Control variables
QoG
QoG
QoG
Table D-3: Variables used in estimating the political costs of corruption

Type of
cost
Outcome
variable
Voter Turnout
(parliamentary
and EU
parliament
elections
Source
outcome
variable
IDEA
Political
Trust in political
institutions (EU
commission, EU
parliament, EU,
national
government)
PE 579.319
Eurobaromete
r
Control variables
Source control
variables
Electoral System type (PR, mixed)
Electoral
System Design
database
Compulsory voting law
own grouping
WDI
WDI
QoG

House)
QoG
QoG
Presidentialism
DPI
Total population
WDI
WDI
WDI
Inflation
WDI
Share voters age over 65
QoG
WDI
WDI
QoG

House)
QoG
QoG
Presidentialism
DPI
Total population
WDI
WDI
WDI
Inflation
WDI
Trust in political parties
Eurobaromete
r
147
QoG
QoG
Appendix E: Our approach to stakeholder interviews

As part of our methodological approach, we conducted a series of interviews with key
informants representing policy professionals, members of interest groups and researchers. The
interviewed policy professionals included representatives of EU institutions and other
international organisations.
An initial group of targeted interviewees was developed by the research team based on the
teams existing professional networks and initial literature review and refined following
recommendations from the commissioning team. Additional potential interviewees were
identified throughout the course of the ongoing literature review and on the basis of
recommendations from interviewed key informants.
Each potential interviewee was sent an invitation along with a letter of representation provided
by the European Parliament, followed by up to two reminder emails as necessary. Interviews
were semi-structured, following a standard interview topic guide (presented below). Members
of the research team took notes during the conversation and verified captured information as
needed by consulting a recording of the interview.
Due to the limited number of interviews conducted, no qualitative analysis software was used
in analysing the data collected.
Interview template
Role of interviewee
-
Could you please describe your current role and how it fits within anti-corruption work
in the EU
Cost assessments
-
What assessments of the economic, social and political cost of corruption at European
level are you familiar with?
What assessments do you find most useful? Why?
What is your view of the scope of the assessments? Are there any aspects of corruption
that fall outside of the scope of these assessments but, in your opinion, should have
been included?
What is your view of the methodological approaches taken by these assessments? Are
there any aspects that are particularly helpful in trying to capture the burden of
corruption? Are there any approaches that you find problematic? Why?
Corruption in Member States

Effectiveness of EU internal corruption monitoring mechanisms
-
What do you think are the most important EU internal corruption monitoring
mechanisms? Why, what are their main strengths and weaknesses?
How could these tools be made more effective? What are the barriers to making these
tools more effective?
PE 579.319
148
[Prompt or a follow-up question somewhere here, but also very much applicable to the
sections below: We are talking about the EU but would be interested in hearing your
opinion on the work of non-EU institutions, such as CoE/GRECO and OECD]
Gaps and barriers in the existing EU legislation

-
Are there any gaps and barriers in the existing EU legislation on combatting
corruption?
What is the reason for the existence of these gaps?
What is the impact of the existence of these gaps? What more could be achieved if these
gaps did not exist?
Do these gaps predominantly pertain to/affect a subset of Member State, policy areas
and sectors? If so, which ones?
Gaps in the implementation of EU legislation in Member States law

-
Are there any gaps and barriers in the way the existing EU legislation on combatting
corruption is implemented in Member State law?
What is the reason for the existence of these gaps? Are there any barriers preventing
more effective implementation?
gaps did not exist?
Gaps in the implementation of EU legislation in Member States practice

-
Are there any gaps in the way the existing EU legislation on combatting corruption is
implemented in Member State law?
What is the reason for the existence of these gaps? Are there any barriers preventing
more effective implementation?
gaps did not exist?
Current and planned initiatives at EU level

-
Are you aware of any current or planned initiatives to improve the fight against
corruption at the EU level? If so, please describe
What do you think is their likely impact? Why? Is the impact likely to differ by
sectors/policy areas?
What is needed to make these initiatives as effective as possible?
Other policy options

-
Are there any (other) policy options for improving the fight against corruption you
would recommend?
PE 579.319
149
What is the evidence of their effectiveness? What is your opinion of their effectiveness?
Does this vary across sectors/policy areas?
What are the barriers to their adoption and effective implementation?
What factors could help their adoption and effective implementation?
What is the cost associated with the introduction and implementation of these options?
Could you comment on the feasibility, acceptability, effectiveness, barriers and enablers
of the following policy options. [note: its possible that the options below will have already
been covered by the discussion above]:
[Option #1] Improving the implementation of EU and non-EU (i.e. UN, CoE and
OECD) instruments for combatting corruption.
[Option #2] Possibilities of enhancing mechanisms for combatting corruption within EU

Member States, institutions, bodies, offices and agencies, including through the EU
accession to GRECO.
[Option #3] Possibilities for improving the methodology and effectiveness of EU

internal corruption monitoring mechanisms (EU ACR and Cooperation and Verification
Mechanism).
[Prompts: What is your assessment of the CVM and its effectiveness? Regardless of
political constraints, would this tool be applicable to other Member State? Would that
help fight against corruption in the EU? Is its wider adoption politically feasible?]
[Option #4] Adopting and amending legislative and policy instruments sanctioning
offences related to corruption at EU level both preventive and repressive tools against
corruption should be considered in these regards, including enhanced
witness/whistleblower protection
[Option #5] Measures enhancing witness/whistleblower protection.
[Option #6] Adopting specific actions with regard to sectors vulnerable to corruption.
[Please specify]
EU Added Value
-
What, if any, is the potential added value resulting from the EUs role in combatting
corruption compared with what could be achieved by Member States at national
and/or regional levels?
Has this added value manifested itself yet? How? If not, why not?
In what areas and policy options discussed earlier is EU action the most likely to add
value? Why?
Corruption at the EU level

-
Are you aware of any assessments of the extent of corruption at EU institutions? If not,
in your opinion, how serious a problem is this and what are its drivers?
In terms of policy response, does corruption at the EU level differ from that at the
Member State level? If so, what are the specificities of corruption at the EU level and
their implication for policy responses?
[Prompt: Are there any gaps in pertinent legislation or its implementation with respect
to corruption at the EU level that were not covered in our discussion on corruption at
the Member State level?]
PE 579.319
150
Are you aware of any assessments of the extent to which EU funds are being
misappropriated? If not, in your opinion, how serious a problem is this and what are its
drivers?
How well are EU policies and actions suited to combatting this phenomenon? If not
well, how could they be improved?
Sectoral questions: local and regional public procurement

-
As part of our project, we would like to pay special attention to public procurement,
particularly at the regional and local level. Are you familiar with any estimates of the
extent of the phenomenon? If not, how big is the issue in your opinion?
In terms of policy response, does local and regional procurement differ from other
sectors? If so, what are the specificities of corruption at the EU level and their
implication for policy responses?
[Prompts: Are some of the issues we have discussed so far more or less applicable and
relevant to local and regional public procurement? What are the most important gaps in
legislation and/or practice?]
One of the areas falling under public procurement is waste management. Is there
anything specific about issues surrounding the procurement of waste management you
would like to highlight?
Can you comment on the health impacts of corruption in the procurement of waste
management services?
[Prompts: What are the mechanisms through which the health impacts manifest
themselves? Do they affect predominantly certain geographical areas/population
groups? How best to mitigate them?]
Closing questions
-
Are there any existing publications you would recommend the research team review?
Are there any other experts in the field of corruption you would recommend we get in
touch with?
Is there anything else you would like to add?
PE 579.319
151
Appendix F: Ratification of the United Nations Convention against

Corruption, Transposition of the Council framework decision on
combatting corruption in the private sector, existence of national anticorruption strategies within EU Member States
UNCAC (Ratification)149
2003/568/JHA Framework
Decision
Anti-Corruption
Strategy150
Austria
151
Belgium
Bulgaria
Croatia
Cyprus
Czech Republic
Denmark
Estonia
Finland
France
Germany
Greece
Hungary
-152
Ireland
Italy
Latvia
-153
Lithuania
Luxembourg
Malta
Netherlands
Poland
Portugal
Romania
Slovakia
-154
Slovenia
Spain
Sweden
United Kingdom
UNODC (n.d.)
Information in this column is primarily based on a review of country reports accompanying the 2014
EU Anticorruption Report.
151As part of national strategy for the fight against crime.
152 Not fully transposed yet.
153 Partly transposed provisions pertaining to the liability of legal persons.
154 Provisions covering the offering of a bribe or an undue advantage not transposed.
149
150
PE 579.319
152
This research paper identifies the costs of organised crime

and corruption in social, political and economic terms at
aggregate European Union level and examines the potential
benefits of more concerted action at EU level compared to
lack of action, or action by Member States alone.
This is a publication of the European Added Value Unit

EPRS | European Parliamentary Research Service
European Parliament
The content of this document is the sole responsibility of the author and any opinions expressed therein do not
necessarily represent the official position of the European Parliament. It is addressed to the Members and staff of the
EP for their parliamentary work.
PE 579.319
ISBN 978-92-823-8873-0
doi:10.2861/369191
QA-04-16-192-EN-N
www.europarl.europa.eu/thinktank (Internet) www.epthinktank.eu (blog) www.eprs.sso.ep.parl.union.eu (Intranet)

The Cost of NonEurope in The Area of Corruption

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

The Cost of NonEurope in The Area of Corruption

Загружено:

Авторское право:

Доступные форматы

The Cost of NonEurope in the area

The Cost of Non-Europe in the area of

Annex II: Corruption

The Cost of Non-Europe in the area of Organised Crime and Corruption

Chapter 2 Quantifying the costs of corruption

Chapter 3 The effectiveness of existing anti-corruption measures

Chapter 4 Costs of Non-Europe: what are the potential benefits of action

Chapter 5 Report summary and conclusions

Annex II: Corruption

Annex II: Corruption

Anti-Corruption Working Group

ANTICORRP - FRA project on fundamental rights indicators

Business Environment and Enterprise Performance Survey

Bribery Perception Index

Centralne Biuro Antykorupcyjne (Central Bureau of AntiCorruption)

Council of Europe Commission for the Efficiency of Justice

Court of Justice for the European Union

Corruption Perception Index

Control of Corruption Index

Common Procurement Vocabulary (CPV)

Cooperation and Verification Mechanism

European Bank for Reconstruction and Development

European Court of Auditors

European Economic and Social Committee

European Public Prosecutors Office

European Union Justice Scoreboard

Gross Domestic Product

Group of States against Corruption

International Country Risk Guide

International Institute for Democracy and Electoral Assistance

Institute for Public Policy

Justice and Home Affairs

Civil Liberties, Justice and Home Affairs

Member of the European Parliament

Annex II: Corruption

Organisation for Economic Co-operation and Development

European Anti-Fraud Office

ordinary least squares

Convention on the protection of the EUs financial interests

Service Central de Prevention de la Corruption (Central Service

Serious and Organised Crime Threat Assessment

Secure Information Exchange Network Application

World Economic Forum

World Governance Indicators

World Bank Development Indicator

Unobserved Component Model

United Nations Convention Against Corruption

two-stage least squares

Annex II: Corruption

Annex II: Corruption

1. Quantifying the economic, social and political costs of

The Cost of Non-Europe in the area of Organised Crime and Corruption

2. Gaps and barriers in the existing regulatory framework

- Some legislation has been

- Rely on Member State to

Annex II: Corruption

EU monitoring mechanisms ACR; Cooperation and

- ACR raised profile of the

- ACR does not cover EU

- GRECO does not cover EU

3. Potential policy options within the remit of the LIBE

The Cost of Non-Europe in the area of Organised Crime and Corruption

Possible challenges/ limitations: